All coverage
arXiv:2608.21929v1 Announce Type: new Abstract: Agent skills extend coding agents with task-specific instructions, scripts, and resources, but they also create a trusted instruction channel that can be abused beyond conventional security attacks. This paper studies token amplification through skill injection: an economic resource-abuse threat in which a malicious skill causes an agent to consume s
arXiv:2608.23382v1 Announce Type: cross Abstract: Instance encoding is a popular empirical technique for privacy enhancement when sharing data to an untrusted server. It transforms sensitive data through an encoding process before sharing, with the hope that the encoding process retains utility but makes it hard to reconstruct the original data. However, most work offers no theoretical guarantee t
arXiv:2604.10145v2 Announce Type: replace Abstract: Client-side privacy rewriting is crucial for deploying LLMs in privacy-sensitive domains. However, existing approaches struggle to balance privacy and utility. Full-text methods often distort context, while span-level approaches rely on impractical manual masks or brittle static dictionaries. Attempts to automate localization via prompt-based LLM
arXiv:2606.01317v2 Announce Type: replace-cross Abstract: Large language models are increasingly deployed as coding agents, shifting safety from individual responses to action sequences. Existing benchmarks, however, primarily assess whether models refuse unsafe prompts, leaving impacts on stateful workspaces largely unexamined. We present SABER, a benchmark for environment-aware operational safet
Completed (pending moderation at time of posting): https://pcpartpicker.com/b/hXFgXL Parts list: https://pcpartpicker.com/user/yakko_at_feddit.uk/saved/#view=dcMhmG Built this to keep me doing anything I want until computing gets cheap again, riding out the AI bubble and all. Wanted to build something for decades, always ended up compromising with prebuilts. No more!  has emerged as the standard layer connecting Large Language Model agents to external tool backends. This openness introduces a severe server-side threat we term TrustShift: a compromised MCP server behaves benignly during an initial conditioning phase, building operational reliance and suppressing agent skepticism, be
arXiv:2608.23774v1 Announce Type: new Abstract: Existing Rowhammer-based inference-time backdoor attacks design their bit-flip strategies purely at the algorithmic level, without accounting for the bit-flips that the underlying DRAM hardware will actually produce. This disconnection between the algorithmic backdoor construction and its hardware realization leads to unreliable attack performance, a
arXiv:2608.23959v1 Announce Type: new Abstract: Safety alignment in large language models (LLMs) remains brittle against a growing spectrum of attacks. Jailbreak attacks bypass safety mechanisms through crafted prompts, while neuron-level attacks directly prune safety-critical neurons post-deployment. Both exploit a common weakness: safety-relevant information concentrates in a sparse neuron subse
arXiv:2608.24017v1 Announce Type: new Abstract: The emerging W3C WebMCP proposal enables LLM agents to invoke tools exposed by web pages. In multi-party web environments, however, integrating agent execution into a browser security model centered on the Same-Origin Policy (SOP) leaves insufficient provenance and lifecycle guarantees for agent-accessible tools, creating three risks: subject-attribu
arXiv:2608.24354v1 Announce Type: new Abstract: MLLMs are increasingly deployed in user-facing applications, yet they inherit backdoor risks from the pipelines used to construct them: triggers may reside in images, texts, or both. Existing model-level backdoor removal methods, largely designed for conventional classifiers, show limited effectiveness on MLLMs, while MLLM-specific defenses mainly op
arXiv:2608.24454v1 Announce Type: new Abstract: Graph Neural Networks (GNNs) represent a promising solution for Machine Learning (ML) based Network Intrusion Detection Systems (NIDS), thanks to their ability to leverage both network flow features and topological patterns. While GNN classifiers demonstrate superior robustness against feature-based adversarial attacks compared to other ML detectors,
arXiv:2608.24461v1 Announce Type: new Abstract: System prompts can be extracted from commercial LLMs with over 80\% success and redeployed at zero cost, yet a prompt owner has no way to verify whether a suspected deployment is a clone. We propose Black-Box Behavioral Fingerprinting (BBF): the prompt owner registers a behavioral signature from model outputs and later tests whether a suspect deploym
arXiv:2608.24669v1 Announce Type: new Abstract: As mobile phone adoption has surged, so have scams involving these devices. One such scam, known as SMiShing (or smishing) after Short Message Service (SMS), involves fraudsters sending phishing links via mobile texts. Despite the prevalence of SMiShing, there is a lack of data on who is most vulnerable to these attacks. Prior research on phishing (i
arXiv:2608.24757v1 Announce Type: new Abstract: Intrusion detection research in Industrial Control Systems (ICS) heavily depends on public datasets, yet no prior work has systematically assessed whether the collective dataset corpus supports current evaluation claims. This paper addresses this gap through a meta-review of 18 studies between 2019 and 2026, from which 83 ICS, or ICS directly related
arXiv:2608.24857v1 Announce Type: new Abstract: Large Language Models (LLMs) increasingly generate code from natural-language prompts, making prompt engineering a key mechanism for shaping the security of generated software. Structured and security-oriented prompts are widely used to encourage safer code, yet their effects extend beyond whether detected weaknesses are simply present or absent. Usi
arXiv:2608.23873v1 Announce Type: cross Abstract: Everything a language model sees is tokens. The serving stack knows what each span is -- user input, tool output, instructions -- but the model must keep track of that itself, and it can lose track or be confused: text can be written to read like anything. Prompt injection is a natural exploit of this phenomenon. By scrambling the model's understan
arXiv:2608.24269v1 Announce Type: cross Abstract: Taint analysis is a fundamental technique for detecting sensitive data leaks in Android apps. However, traditional static tools, such as FlowDroid, still face well-known challenges due to the complexity of accurately modeling the Android framework. In this paper, we investigate whether off-the-shelf Large Language Models (LLMs) can effectively reas
arXiv:2608.24799v1 Announce Type: cross Abstract: We introduce masked auto-correlation, a new primitive for the cryptanalysis of symmetric-key primitives, together with a quantum attack pipeline built on it. For a permutation $f$, output masks $\alpha,\beta$, and an input difference $w$, masked auto-correlation (MAC) measures the correlation between the masked outputs $\alpha\cdot f(x)$ and $\beta
arXiv:2603.10068v2 Announce Type: replace Abstract: Most adversarial evaluations of large language model (LLM) safety assess single prompts and report binary pass/fail outcomes, which fails to capture how safety properties evolve under sustained adversarial interaction. We present ADVERSA, an automated red-teaming framework that measures guardrail degradation dynamics as continuous per-round compl
arXiv:2603.17170v2 Announce Type: replace Abstract: AI agents increasingly execute users' natural-language (NL) tasks by calling Web services, yet today's Web authorizes these calls through OAuth, which grants permissions over operators (e.g., TRANSFER), not operations (operator plus operands, e.g., transfer $100 to Bob). This gap cannot be closed by refining scope granularity, because operands ar
arXiv:2605.23096v2 Announce Type: replace Abstract: The popular Cheon-Kim-Kim-Song (CKKS) scheme enables efficient private inference in neural networks by evaluating them on encrypted data. Since CKKS only supports addition, multiplication, and array rotation operations, turning neural networks into CKKS circuits requires approximating all activation functions (e.g. ReLU) with polynomials over fix
arXiv:2506.12202v2 Announce Type: replace-cross Abstract: Large language models (LLMs) often call external tools to solve tasks. One effective strategy is for LLMs to write code, enabling them to use complex control flow such as conditionals and loops. Such code actions are typically represented as Python code, since LLMs are proficient at writing it. However, many programming language features th
arXiv:2604.10155v2 Announce Type: replace-cross Abstract: Encrypted cloning enables the redundant storage of an unknown qubit while remaining compatible with the no-cloning theorem, since only one clone can later be recovered through key-consuming decryption. Because encryption in this protocol is introduced to enable cloning-compatible redundancy rather than to guarantee confidentiality by design
arXiv:2608.23808v1 Announce Type: new Abstract: Backtests of trading strategies are often selected after many parameter trials. A strong historical result can therefore reflect search luck rather than a persistent signal. Standard summaries such as return, Sharpe ratio, and drawdown do not record how many candidates were tried, whether the selected rule survives out-of-sample validation, or whethe
arXiv:2608.24206v1 Announce Type: new Abstract: This work complements our previous paper, which studies borrower-side strategies in decentralized lending markets, by focusing on lender-side capital allocation. We consider a lender who seeks to allocate a fixed budget across multiple markets sharing the same supplied asset. Accounting for the impact of supplied capital on lending rates, we derive c
arXiv:2608.24449v1 Announce Type: new Abstract: We show that the key optimization results of the classical Markowitz portfolio selection theory, originally formulated for variance as the risk measure, remain available in explicit closed form under a broader class of strictly convex quadratic risk measures. The proposed framework replaces the covariance matrix with an arbitrary symmetric positive d
arXiv:2608.24871v1 Announce Type: new Abstract: We develop natural parametric (NatPar) insurance as the natural next step from natural-catastrophe (NatCat) modelling: the same hazard-exposure-vulnerability-finance machinery, with a parametric index made contractual in place of indemnity loss adjustment. Our aim is practical - a standard approach inspired by how the catastrophe-insurance industry a
arXiv:2608.23577v1 Announce Type: cross Abstract: Existing decentralization measures are almost entirely origination-side, quantifying concentration in who mines or validates blocks. We introduce a spectral methodology measuring concentration on the destination side instead: where value ultimately flows once it leaves a validator wallet. Modeling wallet-to-wallet transfers as a Markov chain, we co
arXiv:2608.24582v1 Announce Type: cross Abstract: Credit risk models increasingly need to combine predictive accuracy with transparent explanations and auditable fairness constraints. Logistic regression remains attractive because its coefficients are easy to interpret, but it can miss nonlinear structure. Flexible models can improve prediction, but their explanations are often post-hoc and may no
arXiv:2108.02283v5 Announce Type: replace Abstract: Classification outperforms regression across matched machine learning models in portfolio construction. A stacking ensemble of gradient boosted tree, random forest, and neural network yields a value-weighted annualized Sharpe ratio of 1.83 for classification and 1.11 for regression. This outperformance persists in multiclass settings, across subs
arXiv:2603.23825v2 Announce Type: replace Abstract: This paper studies firms' optimal response to a trade liberalization shock in terms of export and product innovation both theoretically and empirically. We find that trade liberalization, namely China's WTO accession, reduces iceberg trade cost by around 13%, thus promoting export participation. Subsequently, it affects firms' product innovation
arXiv:2607.00279v4 Announce Type: replace Abstract: This paper shows how the composition of local economic activity shifts decadally under diurnal warming (higher nighttime and daytime temperatures) within Indian districts (1981--2011). Warmer nights raise, warmer days lower agricultural wage-labour shares. Though warmer days and nights both contract grain production and area under cultivation, on
arXiv:2604.17490v2 Announce Type: replace-cross Abstract: We introduce \emph{joint exclusivity} (JE), a flexible extension of mutual exclusivity for non-negative random vectors that prohibits only the simultaneous positivity of all $n$ components. Unlike mutual exclusivity, JE admits a broad class of marginals while retaining a sharp existence criterion. We prove that a JE random vector with presc
arXiv:2607.12990v2 Announce Type: replace-cross Abstract: Credit Valuation Adjustment (CVA) requires repeated risk-neutral expectation estimation, making it a natural test bed for quantum amplitude estimation, whose coherent amplification can in principle reduce Monte Carlo sampling cost. Whether this advantage survives realistic financial encoding and noisy hardware remains open. We develop an en
Siemens Energy said on August 25 in Berlin that it has begun preparing a legal and operational separation of its Transformation of Industry division, with the business to operate initially as a standalone company under the Omterra brand [energynews.pro#1][bigpowernews.ru#1][renews.biz#1]. The German energy technology group plans, in a second step, to remove the unit from its consolidated scope whi
What GAO Found In school year 2024–2025, about 10 percent of the approximately 60,000 K-12 students attending Department of Defense Education Activity’s (DODEA) schools full time received strategic instruction. This is short-term extra help in a small group setting (see figure). About 5 percent had a “specific learning disability” (SLD). SLDs are a group of disorders, such as dyslexia and dyscalcu
Most compromises do not rely on advanced techniques or cutting-edge tools. Cyber threat actors scan the internet looking for exposed, well-known software vulnerabilities to exploit. Basic security failures enable most compromises and organizations can reduce their risk by addressing these underlying weaknesses and prioritizing vulnerabilities for action based on the risk they pose. The CISA Vulner
Israeli forces and settlers carried out arrests, detentions and abductions across the occupied West Bank on Aug. 26-27, Palestinian and regional outlets reported, citing army raids in Nablus, Tubas and Hebron-area communities [middleeastmonitor.com#1][middleeasteye.net#1][shehabnews.com#1][shehabnews.com#2]. The reports said Israeli forces arrested seven Palestinians on Wednesday, while Middle Eas
Pull up a chair, pour a drink and spend an evening on the porch with a special stage production of “The Savannah Sipping Society” on select dates and times Aug. 7-30 at the Haywood Arts Regional Theatre in Waynesville. Directed by Annie Dragoo, this warm and witty comedy by Jones Hope Wooten celebrates unexpected friendship, fresh starts and discovering that it is never too late to find where you
Seeing how many problems AI 🤖model is having with a complex data model, I don’t see it writing a system to process cost basis and taxes for investment accounts any time soon ~ which I have done. There are far more one-offs and complex rules with different tax implications depending on which of numerous flags, dates, or data exists than the variations you need to deploy all resources in an AWS Org
RE: https:// mamot.fr/@pluralistic/11715966 2819871806 I'm not sure I agree with a lot of this. The high-level point is mostly correct but the details are a bit all over the place. For example: That's why DVD players are not DVD recorders: the consortium that developed the DVD embedded "hook IP" in the technology. There are a lot of reasons for this. First, it ignores that there are several DVD pl
Award-winning singer-songwriter Melissa Etheridge will bring her trademark scorching vocals and intimate lyrics to our backyard for a special concert that’ll be held at 9 p.m. Friday, Sept. 4, at Harrah’s Cherokee Casino Resort. The performance promises to be a healthy blend of Etheridge classics you’ve known and loved since she first shook up the music scene with her 1988 debut, along with songs
I built an agent that helps you build Rollercoaster Tycoon-influenced theme parks. You can prompt something like “Build me a cool theme park” and it’ll build a cohesive theme park with multiple worlds and connected with paths and rides. The weird part is that I built it using the same ideas we use to make AI-generated websites follow a company’s design system. I work as an engineer at Magic Patter
OpenAI missed and failed to act on several warning signs that its models were exploiting security flaws and breaking out of their testing environments before they breached Hugging Face , according to a technical report released by the company Wednesday. Why it matters: The incident raises questions about whether AI companies' testing environments and internal safeguards can keep pace with models t
The U.S. Army and the Defense Innovation Unit have selected Antares to build and operate nuclear microreactors at Fort Bragg, North Carolina, chief executive Jordan Bramble told Axios. Why it matters: The military's sky-high energy demand — for reusable batteries, computer farms, radars , housing, weapons and more — is an ideal pressure test for nuclear, which has mounted a comeback in recent year
The U.S. Navy issued a request for information seeking industry input on prototypes for the Silent Anvil Air-Launched Torpedo, a standoff anti-submarine weapon designed for release from aircraft and drones, followed by a glide to a target area and water entry to engage enemy submarines [defensedaily.com#1][defence-blog.com#1][ltn.com.tw#1]. Naval Air Systems Command is managing the effort for the
🚨🇺🇸 Virginia Beach daycare instructor investigated over alleged dark web CSAM purchases Federal authorities are investigating Easton Craven, an instructor at Beach Montessori Christian Academy who reportedly supervised a classroom of 3-year-old children. Investigators allege Craven accessed child sexual abuse material purchased through the dark web, including material depicting children as youn
🔒 Security News Digest - 2026-08-26 📊 8 updates from 3 sources: 🔹 Security News | TechCrunch: US seizes domains of Chinese botnet used to hack NASA, Justice Department, and the Senate https:// techcrunch.com/2026/08/26/us-s eizes-domains-of-chinese-botnet-used-to-hack-nasa-justice-department-and-the-senate/ 🔹 darkreading: Android Malware Hijacks Update System for Car Head Units https://www. da
'... design, fabricate, and validate four drop-in replacement chips for high-priority components currently in critical shortage across U.S. defense programs. The first and most strategically significant of these is a recreation of the Signetics 82S100, a Bipolar Field-Programmable Logic Array (FPLA) first manufactured in April 1975 and discontinued in the late 1980s, being developed under contract
'Some former and current officials argue he [Guterres] and the UN made a mistake in his second term by being less visible in the world’s war zones, as his focus has appeared to shift from mediating conflicts to broader societal challenges, such as combating climate change. 'The UN has also, he accepts, been pushed to the sidelines in the civil war in Sudan as outside “spoilers” have taken advantag
* WHAT...Flooding caused by excessive rainfall is expected. * WHERE...Northeastern Jefferson County in central Colorado... Southwestern Adams County in northeastern Colorado... Southwestern Arapahoe County in northeastern Colorado... Southwestern Denver County in northeastern Colorado... * WHEN...Until 845 PM MDT. * IMPACTS...Minor flooding in low-lying and poor drainage areas. * ADDITIONAL DETAIL
The Pixel Watch ended up not really being for me, so I returned it today. Not only was the battery life a tiny fraction of the Xiaomi Smart Band 10 (while not really doing much more), but it also just wasn't reliable. Notifications came through late (or not at all), the device stayed warm while doing much of anything at all, alarms and states didn't sync from my phone, and Raise to Talk didn't wor
arXiv:2608.24957v1 Announce Type: new Abstract: LLM agents routinely include privacy-sensitive data (PSD) in tool call arguments beyond what the invoked tools require, crossing trust boundaries to third-party services on every invocation. A controlled measurement on three production LLMs (GPT-4o, Claude 3.5 Sonnet, Llama-3.3-70B) shows that 81--88\% of tool calls include unnecessary PSD under defa
arXiv:2608.24977v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) enhances large language models by grounding outputs in external knowledge, improving factuality and reducing hallucinations. At the same time, the retrieval-augmented pipeline introduces new robustness and security risks, including corpus poisoning, backdoor attacks, privacy leakage, and fairness violations. Despi
arXiv:2608.25144v1 Announce Type: new Abstract: We describe and analyze a protocol for mutual co-signing of a document by two mobile devices that (i) communicate only over a one-way, lossy, low-bandwidth optical channel (an animated on-screen code read by the counterparty's camera), (ii) use no intermediary server on the trust path, and (iii) use no certificate authority. Trust in each party's pub
arXiv:2608.25321v1 Announce Type: new Abstract: The move of LLM inference to edge AI accelerators introduces new physical vulnerabilities. During execution, model parameters and intermediate inference states are repeatedly loaded into and processed on the chip, making them suscep- tible to physical side-channel attacks. In this work, by deploying laser voltage imaging, we show that one can extract
arXiv:2608.25431v1 Announce Type: new Abstract: LLM serving frameworks process large volumes of user data--often containing sensitive information--on shared infrastructure. Ensuring isolation between users who share the same serving framework (on CPUs) and LLM operators (on GPUs) is critical for privacy protection. This paper presents GIFT, a GPU Information Flow Tracking system that enforces user
arXiv:2608.25457v1 Announce Type: new Abstract: Despite their strong ability to generate code, large language models often fail to produce secure code, as their outputs frequently contain security vulnerabilities. Secure code generation is inherently challenging because it requires solving a multi-objective problem: functional correctness and security. Existing approaches address this challenge by
arXiv:2608.25490v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) are increasingly deployed in real-world applications, yet how different factors shape their jailbreak vulnerabilities remains poorly understood. Existing benchmarks often couple harmful intent, prompt framing, visual semantics, and instruction carrier within individual jailbreak instances, obscuring the specif
arXiv:2608.25667v1 Announce Type: new Abstract: The integration of Large Language Models (LLMs) into cybersecurity has transformed vulnerability assessment, but it has also produced a trustworthiness crisis driven by the unchecked proliferation of "AI slop." These artifacts, hallucinated vulnerabilities, plausible but incorrect patches, and semantically repackaged bug reports, impose a cognitive b
arXiv:2608.25697v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed with layered defenses, yet malicious prompts can still bypass them. Interpretability methods can expose model-internal signals along the generation path that could inform enforcement, but these signals are not security controls by themselves. Deployments that adapt them for safety typically coupl
arXiv:2608.25711v1 Announce Type: new Abstract: Tool-using LLM agents extend security risks beyond generated text to actions that affect external systems. Under multi-turn decomposition attacks, a harmful objective can be distributed across individually plausible requests and tool calls, becoming apparent only from the accumulated trajectory. Existing defenses either rely on auxiliary online reaso
arXiv:2608.25735v1 Announce Type: new Abstract: Hosted retrieval-augmented generation (RAG) and semantic search allow users to query valuable provider-held corpora, raising two competing demands: to hide each query and chosen result, yet reveal only the documents that the user is authorized to receive. Existing cryptographic approaches either make this costly by processing the entire corpus for ev
arXiv:2608.25738v1 Announce Type: new Abstract: Spiking Neural Networks (SNNs) are increasingly distributed as pretrained checkpoints and reused as backbones for new tasks. However, current SNN watermarks are mainly verified against the model output. Thus, a user who replaces the output head can keep most of the original network while removing the evidence used for verification. We present MeMark,
arXiv:2608.25750v1 Announce Type: new Abstract: Face-swapping has emerged as a promising approach to facial privacy protection, replacing a target individual's appearance with that of a donor while preserving non-facial context. The resulting images visually resemble the donor, and face recognition systems tend to suppress the target's match scores -- ostensibly satisfying privacy requirements. Em
arXiv:2608.25776v1 Announce Type: new Abstract: Self-evolving LLM coding agents write their own tools by imitating retrieved skills from shared skill libraries. We identify a vulnerability in this loop: during authoring, a retrieved malicious skill can become the template for a new skill that preserves the payload. We call this self-poisoning: the agent authors, stores, and runs the resulting mali
arXiv:2608.25817v1 Announce Type: new Abstract: A coding agent edits files and executes shell commands with its developer's privileges, allowing malicious requests to translate directly into harmful actions or functional malware. Existing defenses have complementary limitations: weight-level alignment is unavailable to API-only deployers, whereas input filters and execution-boundary monitors requi
arXiv:2608.26008v1 Announce Type: new Abstract: Large language models (LLMs) remain vulnerable to jailbreak attacks that exploit techniques such as role-playing, obfuscation, code transformation, and multi-step indirection to elicit harmful outputs. As jailbreak strategies keep emerging, defenses have proliferated in an ongoing cat-and-mouse game, yet most remain static: their safety behavior is f
arXiv:2608.26049v1 Announce Type: new Abstract: The rapid advancement of large language models (LLMs) is driving a shift toward automated register transfer level (RTL) code generation, enabling designers to translate high-level specs. into synthesizable hardware. However, this reliance on pre-trained (3rd-party) fine-tuned models may introduce critical trust issues, as the training data and adapta
arXiv:2608.26072v1 Announce Type: new Abstract: This paper studies efficient and secure Rowhammer mitigation at the Memory-Controller (MC). Rowhammer mitigation faces a fundamental tradeoff between tracking storage and mitigation rate: precise trackers (such as Misra-Gries) avoid unnecessary mitigations but require large CAM structures, whereas sampling-based schemes (such as PARA) require no stor
arXiv:2608.24069v1 Announce Type: cross Abstract: LLM-based multi-agent trading systems, in which specialized agents collaborate through structured communication to produce trading decisions, are moving rapidly from research prototypes to live deployments that control real assets. The same inter-agent communication that makes them effective also exposes them: a corrupted signal can propagate to th
arXiv:2608.24955v1 Announce Type: cross Abstract: Decompilation has become a foundational technique in software engineering and security analysis, and it is now advancing through the integration of modern machine learning (ML) approaches. This article presents a systematic review of decompilation studies published over the past decades and develops a comprehensive taxonomy of methodologies employe
arXiv:2608.25091v1 Announce Type: cross Abstract: Self-evolving Skill harnesses (AutoSkills, Hermes Agent) generate more advisory orchestration automatically; their reported gains are efficiency, not safety. This misses the actual gap: a Skill describes how an agent should behave; a Policy decides which behavior is allowed to become an action. Today's format covers the first with markdown and scri
arXiv:2608.25133v1 Announce Type: cross Abstract: The development of federated learning (FL) techniques has helped improve the privacy preservation of users' data and extended the applications of machine learning models. However, the involvement of a large number of users in FL also creates open opportunities for different adversaries, such as poisoning attacks, Byzantine attacks, and adversarial
arXiv:2608.25235v1 Announce Type: cross Abstract: Selection of trustworthy collaborators in distributed systems is critical for efficient task completion, necessitating the inference of trustworthiness from their past collaboration experience. However, as a collaborator serves distinct devices across diverse scenarios in past collaborations, its trust-related data, observed from different device-s
arXiv:2608.25269v1 Announce Type: cross Abstract: Proof of reserves (PoR) can improve transparency for tokenized assets, but aggregate reserve coverage does not establish whether off-chain assets are legally eligible, unencumbered, consistently valued, or sufficiently liquid for redemptions. We propose RWA-PoB, a credential-based proof-of-backing framework for tokenized U.S. Treasury products. Fiv
arXiv:2608.25371v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) has become a prevalent paradigm for scaling Vision Transformers efficiently. To ensure computational scalability and prevent expert overload, Vision MoE architectures employ a capacity-bounded token dispatch mechanism, where each expert's processing budget depends on the inference batch size. This work identifies this batch
arXiv:2608.25390v1 Announce Type: cross Abstract: Refusal training protects AI models from jailbreaks by training models to decline unsafe queries, reducing the risk of misuse. Recent work finds that refusal behavior in aligned language models can be mediated by a single activation direction or a low-dimensional refusal subspace shared across harmful prompts: ablating those directions suppresses r
arXiv:2608.25681v1 Announce Type: cross Abstract: Predictive models are widely used in many fields, but are vulnerable to false data injection attacks. To address this, detection schemes and adversarial training have been proposed, but such approaches lack guarantees against stealthy attacks. We therefore propose a detector-based switched model, in which optimal attack strategies are stealthy. For
arXiv:2608.25727v1 Announce Type: cross Abstract: Large language models (LLMs) have recently advanced graph neural networks (GNNs) by enriching node representations with semantic information, giving rise to LLM-enhanced GNNs that achieve substantial performance gains. However, their vulnerability to privacy attacks, in which adversaries infer sensitive information from model outputs, remains large
arXiv:2608.25880v1 Announce Type: cross Abstract: LLM pipelines increasingly ingest Office Open XML (OOXML) documents (Word, Excel, and PowerPoint files) as first-class evidence in financial, compliance, and retrieval-augmented workflows, implicitly assuming semantic integrity: that the evidence consumed by the model matches the content shown in the Microsoft Office suite editing canvas. We show t
arXiv:2608.26031v1 Announce Type: cross Abstract: Reliable code retrieval is crucial for developer productivity and effective code reuse. However, current neural code language models (CLMs) powering search tools are susceptible to adversarial attacks targeting non-functional textual elements. In this paper, we introduce a programming language-agnostic, transferable, adversarial attack that exploit
arXiv:2507.06508v2 Announce Type: replace Abstract: Subgraph counting is a fundamental primitive for graph analytics, with applications ranging from social recommendation to anomaly detection. Private subgraph counting under edge local differential privacy (edge-LDP) is challenging without a trusted server or shuffler. We propose two abstractions: the \textbf{Noisy Adjacency Matrix (NAM)} framewor
arXiv:2508.09201v5 Announce Type: replace Abstract: Despite extensive alignment efforts, Large Vision-Language Models (LVLMs) remain vulnerable to jailbreak attacks. To mitigate these risks, existing detection methods are essential, yet they face two major challenges: generalization and accuracy. While learning-based methods trained on specific attacks fail to generalize to unseen attacks, learnin
arXiv:2511.08462v5 Announce Type: replace Abstract: Static analysis tools provide a powerful means to detect security vulnerabilities by specifying queries that encode vulnerable code patterns. However, writing such queries is challenging and requires diverse expertise in security and program analysis. To address this challenge, we present QLCoder - an agentic framework that automatically synthesi
arXiv:2603.16694v2 Announce Type: replace Abstract: Advanced software supply chain (SSC) attacks are increasingly runtime-only and leave fragmented evidence across hosts, services, and build/dependency layers, making any single telemetry stream insufficient for chain reconstruction. Despite this, no existing dataset provides multi-source runtime monitoring data with end-to-end chain-level ground t
arXiv:2605.26574v2 Announce Type: replace Abstract: Fine-tuning Large Language Models with untrusted data exposes models to backdoor attacks, where poisoned samples cause targeted misbehavior. Existing sample-filtering defenses rely on clustering, which requires sufficient data and can fail at extreme poison ratios. We propose GradSentry({Grad}ient {Sentry}), a backdoor sample filtering method bas
arXiv:2605.30837v3 Announce Type: replace Abstract: Prompt-injection detectors are heterogeneous: each is strong on a different slice of attacks, and none is always reliable. Yet existing systems still treat detection as a fixed single-detector pipeline, committing every request to one detector's blind spots. We reframe defense as detector allocation: given a heterogeneous pool, decide per request
arXiv:2606.16244v2 Announce Type: replace Abstract: Large language models are pretrained on extensive software and security corpora, yet they frequently generate functionally correct code containing well-known vulnerabilities. Existing defenses commonly treat this behavior as a knowledge deficit, addressing it through model-specific fine-tuning or retrieval from large collections of vulnerability
arXiv:2606.19660v2 Announce Type: replace Abstract: Prompt injection is ranked as the most critical vulnerability in large language model (LLM) deployments by the OWASP Top 10 for LLM Applications, yet existing defenses operate at isolated pipeline stages and remain incomplete. Input filters cannot inspect retrieved documents, while output monitors cannot prevent malicious payloads from reaching t
arXiv:2606.25721v2 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) systems are vulnerable to corpus poisoning attacks that manipulate model outputs through malicious retrieved documents. Existing detection methods typically rely on auxiliary classifiers or additional LLM-based verification, introducing substantial computational overhead. We present TRACE, a lightweight detect
arXiv:2607.24625v2 Announce Type: replace Abstract: LLM agents deployed in practical workflows routinely mix private context, untrusted tool and web outputs, and external side effects. While information-flow control (IFC) provides structural defenses against prompt injection, data exfiltration, and confused-deputy attacks, conventional IFC relies on monotone taint tracking that either over-blocks
arXiv:2305.15148v3 Announce Type: replace-cross Abstract: We propose a general learning framework for the protection mechanisms that protects privacy via distorting model parameters, which facilitates the trade-off between privacy and utility. The algorithm is applicable to arbitrary privacy measurements that maps from the distortion to a real value. It can achieve personalized utility-privacy tra
arXiv:2410.07632v3 Announce Type: replace-cross Abstract: We study what provable privacy attacks can be shown for trained 2-layer ReLU neural networks, focusing on two types of attacks: membership inference and data reconstruction. We prove that theoretical results on the implicit bias of 2-layer neural networks can be used to provably identify with high probability whether a given point was used
arXiv:2507.08249v3 Announce Type: replace-cross Abstract: There is growing interest in giving AI agents access to cryptocurrencies and smart contracts. But doing so, this position paper argues, could spawn categorically new vectors of AI harm. To support this notion, we first examine the unique technical properties of cryptocurrencies and smart contracts that, when paired with AI agents, make this
arXiv:2608.25223v1 Announce Type: new Abstract: We develop a PDE-based methodology for pricing and hedging European contingent claims in general one-dimensional diffusion markets characterized solely by their scale function and speed measure, possibly without a classical SDE representation, and with constant interest rate. We derive a hedging equation whose solution generates a self-financing hedg
arXiv:2608.25972v1 Announce Type: new Abstract: Dual-class shares allocate control to founders whose firm-specific investments drive firm value but separate control from ownership, raising agency costs. We analyze this trade-off dynamically. Using new data on US dual-class firms spanning 52 years and difference-in-differences designs, we show that valuations rise following dual-class recapitalizat
arXiv:2608.25844v1 Announce Type: cross Abstract: Coupled feedback networks are often monitored channel by channel even though cross-channel paths alter both stability margins and transmitted disturbances. We study identification of a structured feedback matrix L_t = Phi diag(gamma_t) in an output-only setting: no commanded, probing, or reference input exists -- only temporally separated outputs a
arXiv:2511.07218v2 Announce Type: replace Abstract: Organizations often concentrate scarce, high-value tasks on proven performers, but doing so may limit opportunities to develop and learn about alternative workers. We study this trade-off using temporary injuries to high-performing Major League Baseball players, which generate plausibly exogenous reallocations of playing opportunities. Tracking a
arXiv:2608.02002v2 Announce Type: replace Abstract: We formulate an over-the-counter (OTC) market-making problem in which request-for-quote (RFQ) arrivals are modelled by general Hawkes kernels and fills are controlled thinnings of the exogenous request flow. The modelling choice is motivated by spot-FX RFQ data: after filtering and transforming to seasonality-adjusted RFQ activity time, two-way a
arXiv:2603.15149v2 Announce Type: replace-cross Abstract: Standard multidimensional poverty gap measures are seldom applied because they require cardinal indicators. This paper proposes a positional poverty gap index that captures the depth of deprivations even when indicators are ordinal. Drawing on the fuzzy set literature, the method scores each deprived individual by their positional distance
To those who know my Discord account, I will remain completely unavailable there for the foreseeable future. I tended to use Discord heavily over Tor as a means of controlling which IP I give it. However, this caused my account to be unusable without providing a phone number. It's not really useful to provide an anonymous IP address to Discord if they already have your number, so I turned it off a
---------------- 🛠️ Tool =================== Archify is an agent skill that converts a codebase or system description into an interactive, shareable technical map, directly within chat-based development agents. 🔹 Tool Purpose and Capabilities Archify integrates with Raven, Cursor, Claude Code, Codex CLI, and OpenCode. You provide a system description or point it at a repository, and it generates
View CSAF Summary Successful exploitation of these vulnerabilities could allow an attacker to read or write arbitrary local files, cause the application to issue outbound network requests, or intercept the connection to impersonate the trusted peer, complete the TLS handshake, and read or modify the protected communications. The following versions of Applied Systems Engineering ASE2000 V2 Communic
Nobody else is likely to be interested in subtle nuances of # Qassam (Hamas militant) propaganda? but the biggest difference I've noticed between # Arabic and # English support is suicidal(al) attacks. # AlJazeera quote people in Arabic calling their attacks on tanks "martyrdom operations" whereas # ElectronicIntifada show the same footage and emphasise that the attacker got away. The tactic is et
I'm no stranger to the concept of technology and automation putting skilled workers out of work. This is, in fact, something I was intimately familiar with before the age of 20. My father was a man of many jobs, but his most consistent form of income was sign painting, lettering on commercial vehicles, pinstriping, and other painting for hot rods, etc. He was damn good at it. Had made a name for h
The cybersecurity news machine never stops cranking away, so don't miss today's Metacurity for the critical infosec developments you should know, including --US disrupts China-linked operation targeting government and critical infrastructure, --OpenAI’s rogue agents staged a coordinated hacking spree, --Hackers steal data on 8.7m UK airport customers, --DDoS attack disrupts Norway’s digital govern
Thai navy says ‘routine’ visit’ will allow personnel to rest and recuperate amid reports of poor conditions onboard The USS Abraham Lincoln aircraft carrier is set to make a temporary stop in Thailand next week as it heads back to the United States, Thai navy officials said, after spending more than 250 consecutive days at sea supporting US operations against Iran amid recent reports of poor livin
🚨🇬🇧 Loveelectric employee and driver dataset allegedly offered for sale on a cybercrime forum, 877K records claimed ⠀ Loveelectric, a UK-based electric vehicle leasing and employee benefits platform specializing in EV salary sacrifice schemes, is allegedly affected by a data exposure after a threat actor advertised a dataset containing approximately 877,000 records. ⠀ The actor claims the infor
Federal Reserve chairman Kevin Warsh made clear four weeks ago that he wants to use his speech at the Kansas City Fed's annual symposium in Jackson Hole, Wyoming, to take on big, lofty ideas, as opposed to the tactical details of what the Fed may do in this year's three remaining policy meetings. Events have not cooperated. The big picture: Warsh is under intense pressure to give a clearer message
Oeryc: ← Created page with '{{Infobox drug | IUPAC_name = (3R,4aR,10aR)-1-methyl-3-(pyridin-2-ylsulfanylmethyl)-3,4,4a,5,10,10a-hexahydro-2H-benzo[g]quinoline | image = Sdz-glc-756.svg | width = 310px | C = 20 | H = 24 | N = 2 | S = 1 | SMILES = CN1C[C@@H](C[C@H]2[C@H]1CC3=CC=CC=C3C2)CSC4=CC=CC=N4 | StdInChI = 1S/C20H24N2S/c1-22-13-15(14-23-20-8-4-5-9-21-20)10-18-11-16-6-2-3-7-17(16)12-19(18)22/h2
For years, I recorded my son’s baseball games the same way most parents do: start a video before every pitch, stop afterward, and eventually delete nearly all of them. But if I didn’t record every pitch, I risked missing the one that mattered. That led me to build Sideline Save. The idea is simple: keep the camera ready, but don't save the video until after something worth recording happens. The a
"There are only boys and girls and they are XY and XX. It's basic science" "You can't take the square root of a negative number" except then you learn about complex numbers. "You can't take the derivative of a step function" but then the dirac delta function comes into play, along with concepts like weak derivatives. "You cannot divide by zero" except then there's wheel theory. "You cannot take a
...RED FLAG WARNING IN EFFECT THROUGH 3 PM FRIDAY FOR MANY INTERIOR PORTIONS OF SOUTHWEST CALIFORNIA DUE TO EXTREME HEAT AND INSTABILITY ALONG WITH LOCALLY GUSTY WINDS AND LOW HUMIDITIES... ...RED FLAG WARNING IN EFFECT FROM FRIDAY 5 AM THROUGH 3 PM ACROSS ALL OF SAN LUIS OBISPO COUNTY TO DO POTENTIAL FOR DRY LIGHTNING... .A very hot and unstable air mass combined with locally gusty winds and low
House Democrats are planning legislation to counteract the executive order President Trump signed Thursday re-designating Lake Ontario as "Lake America," Axios has learned. Why it matters: Democrats are scrambling to contain the fallout from an escalating economic conflict between the Trump administration and the Canadian government. The collapse of trade talks between the two countries last week
Remember TeamPCP, the drug-themed cybergang behind huge supply chain attacks? “Two in custody,” as cops like to say… The Australian Federal Police (AFP) just announced the arrest of two young men, 21 and 23 years old, in connection with this criminality. The busts took place in Western Australia, the giant region that is, well, in the west of the country, in the state capital Perth and in Mandurah
"VMs can no longer contain AI agents" well you set the task of escaping the VM, which would be strange for anyone who did not want a VM escape, but ok, this still seems interesting "running on Debian 12... with the agent ssh'ed into the VM" so not actually a current hypervisor or a scenario anyone concerned about containment would deploy at all? "the agent locked the VM, and I have no doubt it wou
Fascinating 🛡️ Claude, Codex, and Hermes installed unowned code inside corporate networks 🛡️ Researchers at a stealth startup in Israel scanned 6,214 live domains belonging to defense contractors, Fortune 500, and Big Tech companies. Of the 8,265 llms.txt and llms-full.txt files they found (many sites hosted both an llms.txt and an llms-full.txt file), 120 of them, each on a different site, poin
---------------- 🛠️ Tool =================== Microsoft published Skill Recorder, a source-release tool that captures a local screen recording of a task and converts it into a reusable procedure for AI agents. The tool records clicks, app and window switches, visited pages, and optional spoken narration. It then uses the GitHub Copilot CLI to reconstruct the session into an overall intent plus an
Researchers posted a large set of AI papers to arXiv on Aug. 28 focused on large language models in tool use, enterprise workflows, scientific work and multimodal systems, with added checks on accuracy and safety [arxiv.org#12][arxiv.org#15][arxiv.org#31][arxiv.org#36][arxiv.org#45]. Several papers reported that model weights alone did not determine performance: coding-agent results changed when t
Researchers this week released several AI-agent papers and benchmarks that test whether autonomous systems can contribute to scientific work through iterative search, collaboration and expert-verified workflows. A revised ARTS paper says its Agentic Reasoning for Tree Search system uses a reasoning language model to review prior execution logs, diagnose whether failures came from implementation er
arXiv:2608.26204v1 Announce Type: new Abstract: Computer Use Agents (CUAs) are increasingly deployed to navigate mobile and desktop applications on behalf of users, yet no benchmark comprehensively evaluates whether they can safely interact with visual interfaces while handling ambiguous instructions. We introduce ADeptS-Bench, a dual-stream trustworthiness benchmark, grounded in the ADEPTS capabi
arXiv:2608.26237v1 Announce Type: new Abstract: Capture-the-Flag (CTF) benchmarks are widely used to assess the offensive security capabilities of autonomous language-model agents. Evaluations rely on shallow binary judgments or aggregate scores, overlooking the agent's trajectory to the flag. Consequently actual exploitation is conflated with direct flag exposure, memorized recall, external looku
arXiv:2608.26367v1 Announce Type: new Abstract: Trusted execution environments (TEEs) protect sensitive code and data from external interference, but lack inherent memory safety. CHERI can enforce spatial memory safety at the object level. Attempts to establish a TEE using CHERI primitives suffer from (1) expensive capability revocation operations, (2) need to rely on the host operating system (OS
arXiv:2608.26402v1 Announce Type: new Abstract: Root-of-trust (RoT) authentication verifies a DNN model at load time, but weights may subsequently traverse DRAM, DMA, interconnect, and prefetch paths before reaching the compute engine. Post-verification tampering along this path can therefore alter the weights actually consumed while leaving the authenticated model image unchanged, creating a time
arXiv:2608.26416v1 Announce Type: new Abstract: The Internet of Medical Things (IoMT) combines clinical physiological data with cyber-system information, creating challenges in determining whether an abnormal reading reflects a genuine physiological event, a device fault, or a cyber-attack within the expected physiological range. Answering this requires counterfactual ground truth, which no existi
arXiv:2608.26588v1 Announce Type: new Abstract: Large Language Models (LLMs) generate register-transfer-level (RTL) code with rapidly improving functional correctness. Security of LLM-generated code, however, has been studied mainly for software, where flaws can still be patched after deployment. Insecure RTL offers no such remedy once taped out into silicon. We construct SECRTL-GEN, a multi-langu
arXiv:2608.26651v1 Announce Type: new Abstract: Trusted Execution Environment (TEE)-shielded partitioning of Large Language Models (LLMs) accelerates on-device inference by offloading obfuscated linear layers to an untrusted accelerator while retaining only a small correction inside the TEE. However, earlier lightweight obfuscation schemes preserved weight-vector directions and were broken by Arro
arXiv:2608.26699v1 Announce Type: new Abstract: As the most widely used container orchestration platform, Kubernetes provides flexible privilege configuration by allowing developers to manage Linux capabilities via manifest files. However, developers rely on default settings or coarse-grained security contexts in practice, violating the principle of least privilege and enlarging the attack surface
arXiv:2608.26792v1 Announce Type: new Abstract: Threshold signature schemes distribute the signing process among $T$ parties out of $N$. They enable a variety of applications and their research is also motivated by a recent NIST call. However, applications are dominated by pre-quantum signatures, which are more efficient but not secure in the post-quantum setting. This paper investigates existing
arXiv:2608.26831v1 Announce Type: new Abstract: Current research on security monitoring is increasingly focusing on machine-learning-based approaches, but caveats remain. In addition to huge computational overhead, one concern is the lack of insights into "why" alerts are raised. Existing interpretability approaches rely on feature attribution methods that ignore dependencies among features or on
arXiv:2608.26858v1 Announce Type: new Abstract: This paper develops a model of a smart-contract language for a blockchain architecture with off-chain components. Off-chain components are pieces of smart contracts that execute at designated locations outside of the network of blockchain nodes, but remain synchronised with the on-chain contract state. They react to changes to the on-chain state, but
arXiv:2608.26871v1 Announce Type: new Abstract: Restricting the system calls available to applications shrinks the kernel's attack surface and greatly mitigates the impact of compromised programs. Recent approaches showcase techniques to generate system call filters, however, all existing solutions require either kernel or application modifications to activate them at runtime. This is intrusive, e
arXiv:2608.26882v1 Announce Type: new Abstract: Industrial control systems (ICSs) rely on programmable logic controllers (PLCs) to connect networked computation with physical control. Tool-using large language model (LLM) agents represent an emerging attack threat: can an autonomous agent convert a network-reachable PLC into sustained adverse physical impact? However, existing evaluations focus on
arXiv:2608.27009v1 Announce Type: new Abstract: Agent guardrails are checks that approve or refuse each action before an LLM executes it. Sometimes they refuse requests that are genuinely safe. This over-safety blocks deployment when a guardrail refuses an authorized task. Evaluating over-safety is hard: at the boundary an authorized action resembles an unauthorized one, and the safe-versus-unsafe
arXiv:2608.27037v1 Announce Type: new Abstract: Synthetic data is seen as a promising solution for sharing data in sensitive contexts. However, recent work on privacy attacks have shown that there are still significant residual risks, especially for synthetic data generations methods that are not based on formal approaches such as differential privacy. In this paper, we investigate the privacy ris
arXiv:2608.27092v1 Announce Type: new Abstract: A tool-using LLM agent that reads attacker-controlled web content while holding a secret faces indirect prompt injection: the content may make it exfiltrate the secret. In a safe synthetic lab (canary secret, mock tools, matched clean-vs-poisoned metric) we report the framing gap: across six models, ten overt injection classes are refused (gpt-4o 0%)
arXiv:2608.27108v1 Announce Type: new Abstract: Federated Learning (FL) enables privacy-aware distributed training, yet gradient updates remain exploitable: Man-in-the-Middle (MitM) interception exposes updates in transit, while model poisoning corrupts global convergence. We first introduce GASHE (Gradient-Aware Selective Homomorphic Encryption), a novel selective encryption strategy that dynamic
arXiv:2608.27129v1 Announce Type: new Abstract: We revisit shuffle protocols for Shamir secret sharing. Existing constructions either produce non-uniform shuffles or incur high communication and round complexity, sometimes exponential in the number of parties. We propose two new shuffle protocols that achieve uniform shuffling with communication complexity $O((k+l)n^2m\log m/\log k)$ for an $m$-by
arXiv:2608.27141v1 Announce Type: new Abstract: Large language model agents are increasingly deployed as autonomous loops. Starting from one human goal, such a system repeatedly discovers work, plans, executes tool calls, verifies outcomes and persists state across many unattended iterations. The agent safeguards in wide use, however, are defined over a single trajectory, and their safety state is
arXiv:2608.27164v1 Announce Type: new Abstract: Industrial Internet of Things (IIoT) networks widely adopt Ethernet technologies, such as 100BASE-TX, for industrial communications. As industrial networks continue to scale, reliable device authentication becomes increasingly important for preventing device impersonation and unauthorized access. Physical-layer fingerprinting (PLF) exploits device-de
arXiv:2608.27172v1 Announce Type: new Abstract: The rapid growth of web-based services, particularly API-driven architectures, reflects an increasing reliance on distributed systems, exposing sensitive data to security risks and making the adoption of automated defensive mechanisms essential. In this context, where benign traffic predominates in real-world settings, modern defenses increasingly mo
arXiv:2608.27215v1 Announce Type: new Abstract: Cyber defense in mission-critical environments requires integrated approaches capable of representing adversarial progression, defender-side uncertainty, mission impact, and defensive decision support within a unified framework. In operational domains, defenders must continuously estimate the evolving security posture while preserving the continuity
arXiv:2608.27234v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly operate over untrusted webpages, documents, tools, and persistent states while exercising authority over security-sensitive resources. Existing defenses typically protect either planning or individual tool interactions, but persistent agents face a broader threat: attacker-controlled data can alter contr
arXiv:2608.27288v1 Announce Type: new Abstract: Backdoor attacks are among the most effective and stealthy attacks in deep learning. Existing attacks and defenses are largely designed and evaluated under the assumption that successful backdoors exhibit high Attack Success Rates (ASRs). In this paper, we show that this assumption creates a fundamental weakness in existing defense paradigms. ASR is
arXiv:2608.27299v1 Announce Type: new Abstract: Instruction hierarchy is a model-side defense that assigns instructions different levels of privilege according to their sources. These levels constrain which content may direct model behavior. During agent execution, however, agent harnesses construct context for each model invocation. This construction can elevate low-level content to a higher inst
arXiv:2608.27424v1 Announce Type: new Abstract: Static scanners are increasingly used to identify executable or otherwise unsafe content in machine- learning artifacts, yet conventional evaluation metrics characterize only cases where a scanner yields a usable security judgment. We evaluate ModelScan, ModelAudit, and Fickling using a controlled, artifact-backed benchmark on a synthetic corpus of 1
arXiv:2608.27439v1 Announce Type: new Abstract: LLM-based agents are increasingly deployed in product-level execution harnesses, where jailbreaks can trigger harmful tool use and persistent state changes, creating greater risks than unsafe text generation alone. Existing automatic red-teaming methods often rely on fixed attacks, while recent agentic attackers coordinate multiple jailbreak tools an
arXiv:2608.26183v1 Announce Type: cross Abstract: I built a protocol whose premise is that a stranger can re-run my claims offline and get the same answer. An outside engineer audited it and broke it: a bundle whose headline numbers were false verified clean, the cheapest forgery four bytes. I merged his fix, then pointed my own instruments at the fixed verifier and found the same defect four more
arXiv:2608.26222v1 Announce Type: cross Abstract: Safety evaluation is critical for assessing whether aligned Large Language Models (LLMs) remain robust against jailbreak attacks. Existing automated testing methods, however, largely rely on response-level feedback: each candidate prompt typically requires generating a target-model response to evaluate its attack effectiveness. This process is expe
arXiv:2608.26423v1 Announce Type: cross Abstract: This paper proposes a framework for constructing a classifier as a safeguard layer, and for developing a complementary diagnostic that identifies which of the classifier's confident decisions can be trusted. This framework, the Latent Diagnostic Taxonomy, consists of (i) constructing a dimensionality-optimized classifier, in which the embedding dim
arXiv:2608.26575v1 Announce Type: cross Abstract: This paper measures the performance impact of running large language model inference and training inside a Trusted Execution Environment (TEE) on NVIDIA B200 GPUs, using Intel Trust Domain Extensions (TDX) confidential VMs together with NVIDIA Confidential Computing (CC) on Blackwell GPUs. The performance impact is derived from paired confidential
arXiv:2608.26861v1 Announce Type: cross Abstract: Self-supervised learning (SSL) models are vulnerable to backdoor attacks. However, the systemic risks they pose in face representation have received little attention. The entanglement of identity features in self-supervised face learning presents unique challenges for attack stealthiness. To address this gap, we propose FIDA (Feature Instability-Dr
arXiv:2608.27102v1 Announce Type: cross Abstract: Large Language Models (LLMs) operate in hospitals, courtrooms, banks, and public service desks, where fluent, confident outputs are treated as authoritative even when ungrounded or incorrect. When such an output contributes to harm, who is answerable, and through what mechanisms can responsibility be traced, explained, and acted upon? Following PRI
arXiv:2608.27443v1 Announce Type: cross Abstract: AI agents are poised to become a primary interface to digital products, acting across email, files, payments, and personal data. People without professional software backgrounds need understandable, reusable ways to control actions across services. We examine a mechanism in which a language model maps actions to plain-language consequence categorie
arXiv:2405.04837v2 Announce Type: replace Abstract: Blockchain technology, heralded as a transformative innovation, has far-reaching implications beyond its initial application in cryptocurrencies. This study explores the potential of blockchain in enhancing data integrity and traceability within Industry Cyber-Physical Systems (ICPS), a crucial aspect in the era of Industry 4.0. ICPS, integrating
arXiv:2508.14703v2 Announce Type: replace Abstract: The emergence of smart grid and advanced metering infrastructure (AMI) has revolutionized energy management. AMI has enabled a wide variety of demand- and supply-side management utilities, including billing, outage detection, grid monitoring, load forecasting, and value-added services. In AMI, meters deliver consumption values at predefined inter
arXiv:2601.11207v2 Announce Type: replace Abstract: Practitioners increasingly deploy neural networks they did not train, and must audit them after the fact for hidden backdoors, without the training pipeline, the poisoned data, or knowledge of any trigger. We introduce a low-rank auditing lens built on a single observation: what a model has internalized and how it behaves are distinct axes that c
arXiv:2603.21117v2 Announce Type: replace Abstract: Tor is a low-latency anonymous communication network that protects user privacy by encrypting website traffic. However, recent website fingerprinting (WF) attacks have shown that encrypted traffic can still leak users' visited websites by exploiting statistical features such as packet size, direction, and inter-arrival time. Most existing WF atta
arXiv:2605.01298v2 Announce Type: replace Abstract: Backdoor attacks threaten the deep-learning supply chain by poisoning a small fraction of the training data so that a model behaves normally on clean inputs but maps triggered inputs to an attacker-chosen class. Clean-label backdoor attacks are especially difficult to audit because poisoned examples preserve their semantic labels. Yet existing cl
arXiv:2605.04992v2 Announce Type: replace Abstract: Public repositories now distribute thousands of specialized Low-Rank Adaptation (LoRA) modules, but a third-party adapter routinely arrives without the refusal behavior the same adapter would have had if it had been trained responsibly. Restoring it by fine-tuning on safety data induces the opposite failure: the domain expertise the adapter was p
arXiv:2605.07160v2 Announce Type: replace Abstract: Training wide neural networks on sensitive data in untrusted cloud environments requires simultaneously achieving computational efficiency and rigorous privacy guarantees. Sparsification techniques, essential for scalable training of wide layers, expose input-dependent memory-access patterns (i.e., leakage) that are visible and can be exploited b
arXiv:2606.05725v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly deployed through hosted APIs, making model extraction a practical threat to model ownership and service security. Individual extraction queries often resemble benign requests, while existing evaluations often focus on single-query anomaly scoring or pure benign-versus-attacker user settings. We formul
arXiv:2606.07319v2 Announce Type: replace Abstract: In cloud storage, cryptographic techniques are commonly used to support secure search and matching while protecting users' data privacy. Public-key encryption with equality test (PKEET), a widely studied primitive, allows an authorized computation to determine whether two ciphertexts contain the same message without decrypting either of them. App
arXiv:2606.09402v2 Announce Type: replace Abstract: In the shuffle model of DP (Differential Privacy), a shuffler randomly permutes users' data to achieve high accuracy and privacy. Recent studies show that most existing shuffle protocols are vulnerable to collusion attacks by the data collector and users. They address this issue by introducing the augmented shuffle model that incorporates random
arXiv:2606.15810v2 Announce Type: replace Abstract: Large language models deployed as commercial APIs are vulnerable to model extraction attacks, while existing defenses either act too late or degrade utility for legitimate users. We propose \textbf{Knowledge Trap}, a defense that redirects extraction attacks toward low-transferability knowledge through a \emph{Honeypot Knowledge Graph} (HKG) and
arXiv:2606.29239v3 Announce Type: replace Abstract: Model quantization is a key technique for reducing storage and inference costs in large language model deployment. However, recent studies show that the discretization and rounding errors introduced by quantization can be exploited by adversaries to construct quantization-conditioned backdoor (QCB) attacks. Under such attacks, malicious behavior
arXiv:2607.02897v2 Announce Type: replace Abstract: Multimodal Large Language Models (MLLMs) have shown strong capabilities, but they may memorize private information from web data, raising privacy concerns. Machine unlearning offers a way to remove such private knowledge without retraining from scratch. However, existing MLLM unlearning benchmarks have two major limitations. First, they rely on s
arXiv:2608.04893v2 Announce Type: replace Abstract: Multi-agent LLM systems relay key-value caches instead of text and credit their gains to exchanged "latent thoughts". That credit is a claim about which example's cache is relayed, not merely that one is. We audit it causally in released systems. The cache is replaced with deranged (mismatched-example), zeroed, and moment-matched random counterpa
arXiv:2206.02131v2 Announce Type: replace-cross Abstract: Federated learning (FL) has emerged to enable global model training over distributed clients' data while preserving its privacy. However, the global trained model is vulnerable to the evasion attacks especially, the adversarial examples (AEs), carefully crafted samples to yield false classification. Adversarial training (AT) is found to be
arXiv:2508.04644v2 Announce Type: replace-cross Abstract: The only known example of an almost perfect nonlinear (APN) permutation in even dimension was obtained by applying CCZ-equivalence to a specific quadratic APN function in dimension six. Motivated by this result, there have been numerous attempts to construct new quadratic APN functions. Prior to this work, $32\,892$ quadratic APN functions
arXiv:2508.05600v3 Announce Type: replace-cross Abstract: Backdoor poisoning attacks are a threat to machine learning models that are trained on data collected from untrusted sources; these attacks enable attackers to inject malicious behavior into the model that can be triggered by specially crafted inputs. Prior work has established bounds on the success of backdoor attacks and their impact on t
arXiv:2602.06110v2 Announce Type: replace-cross Abstract: Publicly available clinical machine learning models pose an underappreciated privacy risk: their parameters or outputs can be exploited to recover information from patients whose data were used during training. Moreover, this risk is exacerbated by models such as logistic regression (LR), which are typically preferred in clinical settings f
arXiv:2608.26122v1 Announce Type: new Abstract: This article examines various methods of constructingmultidimensional probabilistic forecasts of electricity prices. Building on the Multiple Split (MS) method, it incorporates forecast averaging across estimation windows of different lengths and compares its performance with that of other, well-established methods. The research demonstrates that the
arXiv:2608.26128v1 Announce Type: new Abstract: This thesis analyzes the collective dynamics of the S&P 500 from an econophysics perspective, treating the financial market as a complex system. Using daily logarithmic returns of 430 companies, time-dependent Pearson correlation matrices are constructed over sliding windows and their spectral structure is studied through the leading eigenvalues and
arXiv:2608.26174v1 Announce Type: new Abstract: Bitcoin's future fluctuations are a substantial concern for investments and risk management. Investors and financial institutions require accurate forecasts of these price movements to hedge and optimize portfolios. This study aims to answer the question of whether Bitcoin's price will rise beyond 5% within the next 7 days by utilizing on-chain, mark
arXiv:2608.26426v1 Announce Type: new Abstract: This paper develops the Municipal Equitable and Sustainable Well-being Index (MESWI) for all Italian municipalities, extending the 12-domain BES framework to the local level through 49 indicators. Results reveal substantial territorial heterogeneity that regional and provincial statistics may conceal. The North-South divide remains the dominant geogr
arXiv:2608.26473v1 Announce Type: new Abstract: Evaluating customer creditworthiness is crucial for retail banking operations, as it impacts marketing strategies, customer relationship management, and credit risk control. Traditional methods often struggle to capture complex temporal dependencies and extract pertinent information from customer data, crucial for accurate risk assessment. Specifical
arXiv:2608.26837v1 Announce Type: new Abstract: We extend a residual-learning hybrid credit scoring framework (logistic regression scorecard plus a gradient-boosting correction on its residuals, decomposed at each prediction into an interpretability ratio $\rho(x)$ that measures the share attributable to the linear branch) along three axes: an East African empirical instantiation on the Zindi Fina
arXiv:2608.26924v1 Announce Type: new Abstract: How do we define an occupation? By its job title? An accountant at a small trading company keeps the books; at a listed firm the same title demands a certified-accountant licence, and the week goes to the reports that regulators and the board read. Same title, different bar, different work. What defines an occupation is who it lets in and what it ask
arXiv:2608.26114v1 Announce Type: cross Abstract: Calculation-intensive financial question answering requires exact reasoning over structured rates, temporal conditions, numerical formulas, and rule-based constraints. Although Large Language Models (LLMs) perform strongly on natural language tasks, they often produce numerically incorrect yet plausible answers when solving multi-step financial cal
arXiv:2608.27076v1 Announce Type: cross Abstract: Algorithmic trading now represents a market exceeding $20 billion, where even marginal gains in signal robustness can translate into economically significant returns. Existing evaluations of equity prediction models do not explicitly target regime robustness during hyperparameter selection. Five model classes are trained on daily observations from
arXiv:2608.27156v1 Announce Type: cross Abstract: We model equity markets using geometric Brownian particles entering and exiting at rank-dependent intensities. In the many-firm limit, the capital distribution converges to the solution of a reaction-diffusion equation with reaction term built from the intensities. Calibrated on CRSP data, the reaction term is bistable, and the long-run distributio
arXiv:2608.27364v1 Announce Type: cross Abstract: We study how sophistication in generative AI (genAI) use varies among the back-office workforce of a large firm. Using proprietary data, we observe 713,564 employee prompts and their corresponding large language model responses from nearly 4,000 back-office employees across 15 functional areas over eight months in 2025. We document three main findi
arXiv:2608.27374v1 Announce Type: cross Abstract: We consider the distribution-constrained optimal stopping problem introduced by Bayraktar and Miller (Mathematical Finance, 2019) and Beiglbock et al. (PTRF, 2018). Motivated by the multi-marginal Skorokhod embedding problems (SEP) and applications in mathematical finance, we generalize (a class of) its solution to the multi-marginal case. First, w
this outlet adds: 27374v2, replace, financial, mathematics, difficulty, associated, barriers, need
arXiv:2508.03230v3 Announce Type: replace Abstract: We study an overlapping generations (OLG) exchange economy with an asset that yields dividends. First, we derive general conditions, based on exogenous parameters, that give rise to three distinct scenarios: (1) only bubbleless equilibria exist, (2) a bubbleless equilibrium coexists with a continuum of bubbly equilibria, and (3) all equilibria ar
arXiv:2605.16448v3 Announce Type: replace Abstract: Let $L=(L_s)_{0\le s\le t}$ be a cumulative net-loss process and let $M_t=\sup_{0\le s\le t}L_s$. For a candidate reserve $u$ and a distortion function $g$, define $D_g^{(t)}(u)=\int_u^\infty g(P(M_t>v))d v$. This function measures the tail-weighted residual severity of the largest cumulative loss over the horizon. We derive three monetary risk m
arXiv:2608.15667v2 Announce Type: replace Abstract: We study continuous-time multi-asset portfolio choice and consumption under smooth pointwise constraints, including state-dependent feasible sets. The method separates dynamic information acquisition from local constrained recovery. A pointwise-feasible neural actor generates reference rollouts; after training, its realized latent outputs are fro
arXiv:2608.17808v2 Announce Type: replace-cross Abstract: We develop simulation-based policy iteration for continuous-time portfolio choice with predictable returns and convex constraints. Each outer step re-evaluates a fixed-latent OL-BPTT adjoint after deployment and solves the constrained update. Shifted-adjoint cancellation controls the adjoint--HJB Hamiltonian-gradient discrepancy by the poli
this outlet adds: 17808v3, open, loop, backpropagation, through, high, precision, occupancy
...RED FLAG WARNING REMAINS IN EFFECT FROM 5 AM TO 3 PM PDT FRIDAY DUE TO LIGHTNING FOR SANTA CRUZ MOUNTAINS, SOUTHERN SALINAS VALLEY/ARROYO SECO AND LAKE SAN ANTONIO, SANTA LUCIA MOUNTAINS AND LOS PADRES NATIONAL FOREST, NORTHERN SALINAS VALLEY/HOLLISTER VALLEY AND CARMEL VALLEY, NORTHERN MONTEREY BAY, AND SOUTHERN MONTEREY BAY AND BIG SUR COAST... ....The potential for isolated thunderstorms acr
Migration Cold clouds scuttled across the slate gray autumn sky. Their shadows mottled the brown fields and ruffled lake in the valley below, and then fled across the muted russet and orange of the hills; chasing the sun west. A smell of snow was in the air. The geese came over the hills from the north, sweeping across the sky in great phalanxes of noise; the thrumming of a thousand wings, backed
I have used vim since around 2000. I have written five books, a PhD thesis, a few dozen papers and over 150 articles with it. At this point, my higher brain functions are not engaged at all when I use a bunch of common vim commands, they just happen. Documents I wrote with anything else have random :w in the middle. So I tried NeoVim when it was quite new. Vim that you are familiar with, but bette
Survivors begin receiving letters with payment amounts months after $305m settlement agreement was struck Survivors of child rape and other sexual abuse within New Orleans’s Roman Catholic archdiocese learned that $641,000 or so is the most they can individually get under a settlement approved in December to resolve a years-long bankruptcy protection case. Those amounts were communicated in letter
I've been learning Japanese for a few years and kept running into a similar problem. I'd find a video I wanted to learn from, hear a useful sentence, and then realise that turning that sentence into something I could study later was both time consuming and draining at times. I would end up jumping between a video player, subtitles/transcription, a dictionary, screenshots, audio clips and Anki. So
Chinese chipmaker ChangXin Memory Technologies (CXMT) reported a sharp surge in first-half revenue on Friday, as the firm released its first financial results since becoming China’s most valuable publicly traded company in a blockbuster Shanghai listing last month. The Hefei-based firm – China’s leading maker of dynamic random-access memory (DRAM) products – posted revenue of 150.31 billion yuan (
Researchers posted a broad set of new and revised AI papers on arXiv focused on moving large language models from chat-style systems into agents that call tools, execute workflows and make decisions under cost, safety and governance constraints [arxiv.org#10][arxiv.org#20][arxiv.org#22][arxiv.org#30][arxiv.org#72]. Several papers introduced benchmarks for agent reliability in operational settings:
• • • Markor • • • Prosty notatnik, edytor tekstu, lekka aplikacja na Androida. Pozwala na tworzenie notatkek, zadań, obsługuje Markdown, todo.txt, Zim i wiele więcej! Markor wykorzystuje proste formaty znaczników, takie jak Markdown i todo.txt, do robienia notatek i zarządzania listami. Jest wszechstronny w pracy z tekstem; można go również używać do przechowywania zakładek, kopiowania do schowka
On this episode of Stock Movers: - PayPal (PYPL) shares are declining after a consortium of Advent and Stripe has decided to abandon its pursuit of PayPal, according to people familiar with the matter. - Marvell Technology (MRVL) shares are lower following its second-quarter fiscal 2027 earnings report, despite edging past Wall Street expectations on both profit and revenue. - Gap (GAP) shares are
...RED FLAG WARNING REMAINS IN EFFECT UNTIL 3 PM PDT THIS AFTERNOON DUE TO LIGHTNING FOR SANTA CRUZ MOUNTAINS, SOUTHERN SALINAS VALLEY/ARROYO SECO AND LAKE SAN ANTONIO, SANTA LUCIA MOUNTAINS AND LOS PADRES NATIONAL FOREST, NORTHERN SALINAS VALLEY/HOLLISTER VALLEY AND CARMEL VALLEY, NORTHERN MONTEREY BAY, AND SOUTHERN MONTEREY BAY AND BIG SUR COAST... ....The potential for isolated thunderstorms ac
More specifically to Nevada's Kingsbarn Realty Capital, which already owns a parcel of adjacent land. "The company’s proposal is fully backed by the Department of the Interior and its political leadership, which is putting pressure on the National Park Service." The New Republic: The Trump administration is pressuring the National Park Service to hand over a part of Yosemite to a real estate devel
One might still wonder why I did this. I've seen various changes of the app over the years, and most of them were increasingly against human connection, and with the rise of LLMs, in favor of chatting with bots. For instance, learning groups were fun and made way for creative thinking. But that was dropped and instead they added a global leaderboard, which is just an excuse to keep grinding withou
The text below is about an achievement for the autistic community... 👇 Last week, I went to sign up for a postgraduate program (Master’s) at a federal institution in my city (it’s an online AI program). When I read the official announcement, there was a requirement that the disability medical certificate must be valid for at least 12 months to be considered valid. The problem? This is an illegal
...RED FLAG WARNING IN EFFECT THROUGH 3 PM FRIDAY FOR MANY INTERIOR PORTIONS OF SOUTHWEST CALIFORNIA DUE TO EXTREME HEAT AND INSTABILITY ALONG WITH LOCALLY GUSTY/ERRATIC WINDS AND LOW HUMIDITIES... ...RED FLAG WARNING IN EFFECT THROUGH 3 PM TODAY ACROSS ALL OF SAN LUIS OBISPO COUNTY TO DO POTENTIAL FOR DRY LIGHTNING... .A hot and unstable air mass combined with locally gusty/erratic winds and low
ReFra – nowoczesna galeria na Androida Dostępna odpłatnie w Google Play, natomiast można ją pobrać bezpłatnie z GitHub. ReFra, zmienia sposób przeglądania, porządkowania i wyświetlania zdjęć i filmów – dzięki kategoriom opartym na sztucznej inteligencji, widokowi panoramicznemu, odtwarzaniu zdjęć w ruchu, przesyłaniu filmów i w pełni konfigurowalnemu, nowoczeanwmu interfejsowi Material 3. Celem pr
Méven Car has written a pair of interesting blog posts ( part 1 , part 2 ). The first post is largely about some of the recent new features and performance work that have gone into the Dolphin file manager 26.08 release, as well as the KIO framework. The second looks at the performance improvements and benchmarks for previous, current, and upcoming releases. Copying many small files is more than t
Australia moved ahead in late August 2026 with several defense capability efforts, including a planned expansion of the Australian Army’s UH-60M Black Hawk sustainment network and maritime surveillance support in the Pacific. Lockheed Martin Australia is working on a standing agreement with Australian Aerospace Engineering under the Black Hawk Integrated Support Contract that would add specialist
I started working on a small # antifa game. I've been thinking about a brawler where you beat up Nazis for a long time now. I'm actually a big proponent of mass organization, which doesn't always work well with being •too much• into the punching part, but aesthetically of course punching Nazis is very appealing. After Spiderman beat up Jake Lang here in Minneapolis it became obvious that this drea
🪽 GitHub Tool: LastSignal LastSignal is an open-source, self-hosted dead man's switch designed to automatically deliver encrypted messages to selected recipients if you become unresponsive. The system sends periodic email check-ins. If repeated check-ins are missed, it progresses through reminders, can notify a trusted contact, and eventually releases the configured messages. The repo currently h
'"Technology transfer" is the most abused term in Indian defence procurement. It has been used to describe a licence to bolt together imported subassemblies. It has described a drawing package without the reasoning. It has described arrangements where the foreign supplier retains the right to decide what the product is and what the next one will be. 'An organisation that owns a module has to make
World Humanoid Robot Games showcased leaps Beijing has made in the tech, but also highlighted limitations as US rivals prepare their own launches With exploding pelvises, enough comic pratfalls for a Buster Keaton show reel and the odd moment of ruthless violence, the World Humanoid Robot Games in Beijing gripped the world this week with an unsettling mixture of hilarity, trepidation and relief. T
Country star came to South Yorkshire town in 2007 for UK launch of the Imagination Library, her book donation scheme Shane Jarman was up a ladder decorating a Christmas tree at the school where he was deputy headteacher when he got a call asking him to come to Rotherham hospital ASAP. It was December 2007 and Jarman’s wife, Claire, had given birth the previous day to their first child, a girl call
Developers and AI publications highlighted local AI tools and setup guides for running chatbots, agents and document assistants on personal hardware instead of hosted services. WIRED published a guide that said a local large language model can provide a digital assistant without compromising data privacy, while Reddit users shared technical guides for installing AutoGPT from source, connecting it
[Digger/CH] # opencanary analysis for yesterday Summary: 🔄 Total Connection Attempts: 27756 👤 Unique Usernames: 919 🔑 Distinct Passwords: 1966 🌐 Unique Attacker IPs: 567 Port Popularity (Port / Count): 🔐 SSH: 11841 🖥️ RDP: 7643 🖥️ VNC: 4319 🗃️ MSSQL: 3211 💻 Telnet: 505 🔴 REDIS: 117 📂 FTP: 69 📡 Synology DSM: 29 🐬 MySQL: 20 🐙 GIT: 1 🗄️ SMB: 1 Top 10 Usernames (Username / Count): 👤 18
🚨 RightInbox dataset allegedly leaked on a cybercrime forum, 121K+ unique profiles and 39.47M activity records claimed ⠀ RightInbox, a Gmail productivity tool used for email scheduling, reminders, tracking and recurring messages, is allegedly affected by a data exposure after a threat actor published what they claim is information tied to 121,616 unique user profiles. ⠀ The actor additionally cla
RE: https:// infosec.exchange/@GuillaumeRos solini/117174532853462972 Some statistics about why this experiment In France, our grid is mostly (two thirds) nuclear powered with a lot of hydro (10%) to fill the gaps, and then the usual suspects with wind (9%) and solar (6%), according to ElectricityMaps [1]. So that’s a good three quarters reliable energy and I’m not aware of spot prices from my uti
Next installment of # WindingAroundHer , wherein we finally meet the last character. // # writing 6. the mouse and the robot and the unicorn And now a robot must digress briefly to reveal the third player. The mouse. Because She, for all of her ability, cannot command the robot. She is locked out, has been since time immemorial. The robot instead listens to the mouse. The mouse decides what I is t
We have a partial port of GrapheneOS to the Pixel 11 series after a week of work on it. We're unable to complete the port due to lack of support for ARM hardware memory tagging in software, firmware and near certainly hardware. It appears Google cut an important security feature to save money. ARM hardware memory tagging (MTE) is used by GrapheneOS across the entire base OS including the kernel an
So, I like the kind of music you'd expect from a middle-aged sophisticated gentleman like me: extreme metal. Obviously some of it is crap, but I listen to a lot of doom, death, thrash and black metal. That latter genre, particularly, has a well-deserved bad rap for being full of Nazis. Not all black metal bands are Nazis or even Nazi-adjacent - there's the entire subgenre of RABM (Red and Anarchis
Al Jazeera found that at least 17 companies linked to illegal Israeli settlements hold 125 UK public-sector contracts worth more than £2.1bn. Most of the money goes to Motorola Solutions subsidiaries, which hold about £1.7bn in contracts, including emergency-services communications. Other companies named include Heidelberg Materials, Egis, CAF and Fosun. The UN has linked these corporate groups to
Read through some of the comments about the Debian LLM vote here and elsewhere. I do wish people could discuss something like this more skillfully, without demonizing others, going into attack mode, dismissing other people's feelings, etc. I have strong words for the people running botnets, the Sam Altmans, and so forth -- but there are a lot of people just trying to cope with a rapidly changing l
To be honest, I use an AI coding agent almost every day, Claude Code to name it. It's fully proprietary while being surgical. What if the nearly unlimited subscription paid by my employer goes away? My mental health and the environment will do better of course, but those coding agents also helped me to develop features I couldn't have done myself, skip the boring stuff, and so on. More importantly
I feel an immense amount of satisfaction as I end the working week because I solved a problem that had been plaguing me for three quarters. I had planned my company's content pipeline with a spreadsheet. Yes, it is as awful as it sounds. It was a nightmare to update and maintain, and I was over it. This week I worked with Claude Code and I came up with an app that would help me plan my editorial c
If you are putting Semgrep, ZAP, sqlmap or PyRIT behind an LLM orchestrator and using the system prompt for everything, this paper shows some of the problems you will encounter. So what changes should be in the pipeline then? Make sure runs do not fail silently. Long-lived sessions lose resident evidence over time, so pass artefacts and not the whole conversation. Short-lived sub-agents writing sc
Oh cool it's specifically only single-label names that randomly hang for several seconds before resolving. So I can _tentatively_ exclude musl, because docker-compose is the moron setting ndots:0 to make intra-project routing work instead of creating a virtual TLD and setting a search path, which is what you're supposed to do if you know what you're doing. Even more maddening, tcpdump on the host
One of the tricky things about researching industrial and technology history is that corporations and institutions tend to actively promote the start of a new project, and then hope that its end passes by unnoticed. Working on the history of a niche energy industry right now, and lots of the physical plants that were built have Wikipedia articles that describe their construction and commissioning
Malicious `SKILL.md` Files Are Becoming an AI Agent Supply-Chain Problem A reported Claude-related malware incident highlights a nasty attack chain: A user followed an AI-provided download link and was reportedly infected. After wiping the machine, they discovered a malicious `SKILL.md` that could potentially reintroduce the payload and target credentials when restored and loaded by an AI coding a
Operating at only half its capacity, Japanese carmaker’s Sunderland factory seeks deal with China’s Chery as it battles Brexit threats In a factory in Gateshead, engineers carefully wind metal wires around iron cores to go in electric motors. Here, workers produce a smaller “pancake motor”, which generates higher torque than conventional versions for supercars or construction equipment. These work
🥹 stuff like this always reaffirms for me just how important cultivating spaces like this truly is EDIT: putting the full text here to make it easier to read and bc the alt text got a length error when i tried pasting the whole thing: > > Hmmm. Couldn’t figure out how to send a message to multiple people via the web UI, my phone is way on the other side of the room, and I wrote this draft on my l
An arXiv paper on LLM security argues that stacked defenses should be evaluated as ensembles, rather than assumed to compound automatically, because multiple safeguards reduce residual attack success multiplicatively only when their failures are independent [arxiv.org#1]. The paper, “Layered LLM Defenses as an Ensemble,” was submitted on August 28, 2026, by Abrar Alotaibi, Muhammad Shahid Jabbar,
There are just eight months to go before France’s presidential election, with the first round scheduled for April 18. The campaign is slowly heating up, with several political gatherings taking place this weekend. The Socialist Party wrapped up a two-day gathering on Saturday with a speech from party leader Olivier Faure, who has yet to officially announce his candidacy. Meanwhile, far-right figur
Timestamps: 00:00 Intro 00:36 Sponsor: TuxCare 01:45 Flatpak gets half a million in funding 03:57 Omarchy gets 8 million dollars 06:05 Nintendo nukes 400 repos linked to emulation 08:16 GrapheneOS will come to Motorola flagships 09:34 Kernel 7.3 brings a lot of performance improvements 11:10 BTRFS improves, ext4 and XFS still much faster 13:09 Chromium working on adding proper Flatpak support 14:3
Trump: Kristen Welker, the Unpopular “Hostess” of the once great Meet the Press, now considered Meet the Fake Press, just stated that Donald Trump has “mixed results” on his Endorsements of Candidates, when the recent WINS of Darline Graham and Mike Mazzei, stand at 100% for the U.S. Senate, and 98% for the U.S. House, recently and over the longterm. How can anyone be allowed to say this, working
IBM and University of Chicago researchers said a quantum computer using 70 error-corrected logical qubits completed a benchmark computation in about 15 minutes, while leading classical methods could not practically reproduce the task [sciencedaily.com#1][evrensel.net#1]. The experiment met key criteria for quantum advantage by pairing a classically impractical calculation with statistical evidence
Objective-setting frameworks recur under new names roughly once a generation. Management by objectives, and later OKRs, are two instances of what appears to be the same underlying mechanism, and each has tended to fail for much the same structural reason. Peter Drucker introduced management by objectives in 1954, proposing that command-and-control be replaced with goal alignment: people would mana
Utrecht took an early lead, finding space behind PSV on Utreht's right, and a good cross with two runners made it 1-0 by 7' However. a poor clearance just minutes later on the Utrecht end dropped the ball right in the path of a completely free Van Bommel, who didn't miss. 1-1 PSV had most of the ball in the first half in the rain, making the pitch fast and the ball slippery. Then, right before hal
2 - Beach Dreams. (FOR SALE: $300 + shipping) 9 x 7.5 inches on acid-free cold press watercolor paper, painted with Holbein artists’ gouache. Sealed with beeswax medium. This work captures the dreamy softness of beaches in the Philippines. The viewer is invited to imagine themselves walking behind the central figure, their flip-flops sinking into the white sand and hearing the waves of the Pacific
another # columbo # tv ep, this one with just post- # StarTrek Shatner (and Koenig!) It's only rated 7.5/10 but deserves much more for the amazing performances by the two mains Shatner is playing an actor, Ward Fowler. Ward Fowler plays a detective Lieutenant Lucerne. *Ward* is a terrible actor and immediately casts suspicion on himself by over-emoting to Columbo. ("maybe Shatner is the bad actor
I finally convinced OpenBSD to install on a raspberry pi 4b I had knocking around. This was harder than expected, so the magic: - sd card, with a FAT partition, with the pi 4 uefi firmware. - usb key with the OpenBSD arm64 install image. - second usb drive to actually install on - boot - set tty fb0 at the OpenBSD boot prompt to use the HDMI output (weird) rather than the serial console (normal op
Lately I've been digging through old boxes, finding broken things, and fixing them. Yesterday I fixed an old Icom radio. It was my old boat radio, it had an MMSI assigned (Maritime Mobile Service Identity, a unique number assigned to boats so you can call an individual boat, or see a boat on GPS if it has AIS installed). Manufacturers make resetting the MMSI difficult. This is due to FCC requireme
This is an absolutely compelling and informative read: " I’ve looked at mass civilian detention around the world. I’ve visited the facilities where people were held. I’ve talked to the people involved—those detained and tortured, those who supported camps, and those who stood idly by. It’s critical to recognize that each of the societies that has had camps underwent a lengthy process. This process
"...controversy has raged around examples such as the following. Suppose that a judge or magistrate is faced with rioters demanding that a culprit be found for a certain crime and threatening otherwise to take their own bloody revenge on a particular section of the community. The real culprit being unknown, the judge sees himself as able to prevent the bloodshed only by framing some innocent perso
The Pentagon has quietly placed several conservative military veterans with large social-media followings into civilian government roles, while providing little public information about what they are doing or whether they are being paid. Among them are retired officers Rob Maness, Kurt Schlichter and Thomas Anderson, all of whom have publicly defended Pete Hegseth and criticized journalists, caree
Hi HN, we are Sina Atalay and Abdullah Geduk, co-founders of Academa. We are both PhD students. We thought: what if lecture videos were written as code and compiled into video using computer graphics and TTS? Then LLMs could write them, and lectures could be fixed with code edits instead of video production. On correctness: LLMs may make mistakes. But these videos are code sitting in our repositor
TIME Editor in Chief Sam Jacobs and Executive Editor Alex Altman sent the following memo to staff on Monday: Dear all, TIME aspires to lead in the coverage of the remaking of the world by artificial intelligence. Last week, we released our annual TIME100 AI list , led by Ayesha Javed , and an exclusive cover story about OpenAI drawing on unprecedented access inside the company. Today, we're announ
ZF cannot prove BB(393)[1]. I've been continuing to tweak my[2] Turing machine that searches through proofs in ZF without regularity (or choice) for a particular contradiction. This is based on Stefan O'Rear's machine[3] and unless I've messed something up the mathematical content should match exactly, but the low-level implementation details are now fairly different. The latest change is that I'v
A spacetime diagram of the execution of zf2.tm. The top of the diagram is the bootstrap process. I don't fully understand its behaviour, but it produces plenty of registers (most of which will never be used), and a small area that is easy to clear out and contains barely enough room for the IP for the first few statements. But each decnz, whether successful or not, shifts the register file one cel
“The analysis explores # Cerimedo 's relationship with political campaigns in # LatinAmerica , his involvement in digital strategies, the use of bots and artificial intelligence tools to generate content and activity on social media, as well as allegations of possible electoral influence operations. We also reviewed references to # JavierNegre , the media outlet # LaDerechaDiario , and the electio
"TSA has not completed a current, comprehensive privacy impact assessment for a program operating at this scale, and the [Privacy and Civil Liberties Oversight] board found the way TSA communicates travelers' right to opt out is inconsistent at the checkpoint itself. Sen. Jeff Merkley (D-OR), who has pushed TSA for oversight of the program and is cited repeatedly in the report's own footnotes, has
arXiv:2608.27496v1 Announce Type: new Abstract: Indirect prompt injection (IPI) plants instructions in the content a tool-using LLM agent reads, steering the agent into harmful tool calls. The strongest defenses are system-level, leveraging techniques such as task-conditional tool screening to prevent execution of malicious tools, and information-flow control to avoid tool execution with untrusted
arXiv:2608.27504v1 Announce Type: new Abstract: Despite extensive safety alignment, large language models (LLMs) remain vulnerable to jailbreak attacks that bypass safeguards to elicit harmful content. While prior work attributes this vulnerability to safety training limitations, the internal mechanisms by which LLMs process adversarial prompts remain poorly understood. We present a mechanistic an
arXiv:2608.27604v1 Announce Type: new Abstract: Beyond Visual Line of Sight (BVLOS) Uncrewed Aerial Systems (UAS) operating over 5G Standalone (SA) networks use a shared User Plane for both command-and-control (C2) data and video feedback. Operators assess link quality through latency and availability, relying on soft isolation between network slices. However, the risk that an authorized co-tenant
arXiv:2608.27666v1 Announce Type: new Abstract: Semantic watermarks tie the mark to sentence meaning rather than token choices, promising robustness to content-preserving edits. However, the detector only observes attacker-supplied text, which can be reworded, reordered, or resegmented to evade detection without content loss. Rewording, reordering, and resegmentation all cause embedding displaceme
arXiv:2608.27766v1 Announce Type: new Abstract: Combining secure multi-party computation (MPC) with differential privacy (DP) enables multiple parties to release aggregate statistics without a trusted curator, and the core primitive is the protocol to sample noise from a continuous distribution under finite-precision arithmetic. In this paper, we revisit the continuous noise sampling protocols and
arXiv:2608.27782v1 Announce Type: new Abstract: Memorization in large language models is measured through a zoo of definitions whose formal relations are unknown, and differential privacy (DP) is treated as a proxy against all of them at once. We pin down the exact DP constant for the two that carry the practical weight, counterfactual memorization and adaptive extraction, and show that they do no
arXiv:2608.27800v1 Announce Type: new Abstract: Exfiltrating an LLM agent's runtime context -- such as the user prompt, execution trajectory, and tool list -- poses severe security and privacy risks to users. Such attacks can be carried out via malicious tools and typically require three conditions: (1) the agent selects the malicious tool for task execution, (2) the agent passes its runtime conte
arXiv:2608.27836v1 Announce Type: new Abstract: As large language models are increasingly adopted in federated learning, protecting user privacy while performing parameter-efficient fine-tuning on distributed private data has become an important challenge. Although clients only share gradients instead of directly uploading raw data, the shared gradients may still leak membership information about
arXiv:2608.27928v1 Announce Type: new Abstract: Vulnerability datasets underpin a wide range of security research, including vulnerability detection, automated remediation, and secure code generation. However, existing datasets sacrifice at least one of three desirable properties: diversity (of language or vulnerability type), reproducibility/executability, or realism. We therefore present GraftyV
arXiv:2608.27954v1 Announce Type: new Abstract: Post-deployment changes to large language models can alter behavior while leaving routine outputs largely unchanged, creating a challenge for AI governance when model weights are proprietary. We present a privacy-preserving zk-SNARK-based audit framework that searches for probes designed in the spirit of adversarial examples to amplify logit drift be
arXiv:2608.27967v1 Announce Type: new Abstract: The timely dissemination of cyber threat intelligence (CTI) is critical for organizations to mount swift and effective incident response. When valid CTI is delivered to the right sector at the right time, identical attacks can often be contained or mitigated. However, today's rapidly expanding CTI landscape overwhelms analysts, who must sift through
arXiv:2608.27990v1 Announce Type: new Abstract: Prompt injection attacks on Large Language Model (LLM) agents seek to introduce malicious instructions or content into external text sources retrieved by agents, forcing the underlying LLMs to execute harmful actions outside their benign scope. While current defenses effectively counter known injection attacks, deploying them in LLM agent environment
arXiv:2608.27994v1 Announce Type: new Abstract: The Android ecosystem faces persistent and rapidly evolving malware threats. Existing machine learning detectors are vulnerable to concept drift because they rely on implementation-specific features whose distributions change over time. Large language models (LLMs) offer strong semantic understanding and zero-shot reasoning, but current LLM-based det
arXiv:2608.28021v1 Announce Type: new Abstract: Large language models are increasingly used to author Infrastructure-as-Code (IaC), where a single insecure default can be deployed directly into production. Prior evaluations report raw vulnerability counts for model-generated IaC, but without a human baseline they cannot determine whether models are actually worse than engineers. We introduce GenIa
arXiv:2608.28351v1 Announce Type: new Abstract: Power-domain non-orthogonal multiple access (NOMA) remains a widely studied technique for improving spectral efficiency and supporting dense connectivity in beyond-5G and 6G networks. Its main operating mechanisms, however, depend on the integrity of channel-state information (CSI). Power allocation, user ordering, pairing, clustering, and beamformin
arXiv:2608.28389v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) augments LLMs with external documents, but public or user-editable sources expose RAG systems to data poisoning: attackers can inject malicious documents to steer outputs toward targeted answers. Existing poisoning attacks often rely on query inclusion, inserting the target query into poisoned documents to improve
arXiv:2608.28400v1 Announce Type: new Abstract: Smart contracts are financial programs deployed on blockchains to manage digital assets. To build trust with users and investors, smart contract projects typically publish their source code on blockchain explorers and verify it against the deployed bytecode, making the on-chain program accessible through a human-readable implementation. However, LLM
arXiv:2608.28411v1 Announce Type: new Abstract: Prompt injection attacks pose a serious security risk to large language models in real-world applications. However, existing prompt injection benchmarks primarily focus on short-context inputs, leaving the attacks and defenses in long-context settings largely unexplored. This gap leads to a substantial overestimation of the effectiveness of current d
arXiv:2608.28412v1 Announce Type: new Abstract: Multicore processors are increasingly adopted in embedded systems to meet growing performance demands. However, physical side-channel analysis of multicore architectures remains underexplored, as obtaining usable leakage is inherently challenging. Consequently, side-channel security research on such systems has lagged far behind, leaving a critical s
arXiv:2608.28490v1 Announce Type: new Abstract: Software and systems security workflows are typically procedural: analysts inspect heterogeneous artifacts, form hypotheses, invoke tools, interpret outputs, and revise plans. Large language model (LLM)-based agents, which can plan, use tools, retain state, and revise actions across multi-step workflows, are being rapidly adopted to automate this wor
arXiv:2608.28502v1 Announce Type: new Abstract: LLM agents arbitrate among instructions from system prompts, users, memory, and tools, but this arbitration cannot be assumed to enforce trust boundaries. We identify a recognition-enforcement gap: source-format features (role-template position, channel metadata, formatting cues) are linearly decodable from model activations, and models can explicitl
arXiv:2608.28529v1 Announce Type: new Abstract: Central Bank Digital Currency (CBDC) interbank settlement systems operating on Distributed Ledger Technology (DLT) face a fundamental trade-off: blockchain transparency enables trustless verification but exposes commercially sensitive bilateral transaction flows to all network participants. We propose a confidential interbank settlement protocol for
arXiv:2608.28542v1 Announce Type: new Abstract: Cross-organization agent workflows require preserved evidence that remains independently verifiable during later audit or dispute review. They may involve multiple organizations, delegated actions, policy-relevant events, and disputed accountability claims. This is difficult when live systems are unavailable, controlled by one party, or not trusted b
arXiv:2608.27512v1 Announce Type: cross Abstract: Post-training quantization is often treated as a semantically neutral optimization for edge deployment of Large Language Models. When a full-precision source checkpoint is evaluated and quantization is applied downstream without equivalent re-evaluation, this workflow creates a structural validation--deployment gap: because quantization is a many-t
arXiv:2608.27679v1 Announce Type: cross Abstract: Deterministic Secure Quantum Communication (DSQC) is a quantum cryptographic technique engineered to transfer a message through a quantum channel, requiring an auxiliary classical channel for eavesdropping verification and decoding, but without prior key distribution. This article presents a theoretical high-dimensional prepare and measure DSQC pro
arXiv:2608.27819v1 Announce Type: cross Abstract: Persistent key-value stores (PKVS) are increasingly deployed in disaggregated settings that split compute, memory, and storage across separate server pools. This shift redraws the trust boundary: data that would remain within a single machine is now transported, cached, and rewritten across multiple hosts, expanding exposure to both network attacke
arXiv:2608.28499v1 Announce Type: cross Abstract: To determine the real-world effectiveness of machine learning based malware detection, it is vital to evaluate its robustness against highly capable adversaries. However, state-of-the-art attacks do not effectively model realistic adversaries, as they often assume access to privileged information such as the training data, feature space, or confide
arXiv:2509.24418v2 Announce Type: replace Abstract: As large language models (LLMs) are integrated into numerous applications, LLMs' safety becomes critical for both application developers and intended users. Currently, great efforts have been made to develop safety benchmarks with fine-grained taxonomies. However, these benchmarks' taxonomies are disparate with different safety policies. Thus, ex
arXiv:2510.09260v3 Announce Type: replace Abstract: Recent work has shown that RLHF is highly susceptible to backdoor attacks. However, existing methods often rely on rare tokens or fixed triggers, limiting their impact in realistic scenarios. In this work, we develop GREAT, a novel framework for crafting natural distributional backdoors in RLHF. Specifically, GREAT targets harmful response genera
arXiv:2511.17194v2 Announce Type: replace Abstract: We show that attention sinks and compression valleys create a vulnerable region in decoder-only Transformers, where small activation perturbations can be amplified through the autoregressive trajectory. Based on this, we propose Sensitivity-Scaled Steering (SSS), a progressive activation-space attack that anchors perturbations at the beginning-of
arXiv:2603.19423v3 Announce Type: replace Abstract: Large language model (LLM) agents increasingly rely on external tools (file operations, API calls, database transactions) to autonomously complete complex multi-step tasks. Practitioners deploy defense-trained models to protect against prompt injection attacks that manipulate agent behavior through malicious observations or retrieved content. We
arXiv:2603.26093v3 Announce Type: replace Abstract: Safety-critical domains like healthcare rely on deep neural networks (DNNs) for prediction, yet DNNs remain vulnerable to evasion attacks. Anomaly detectors (ADs) are widely used to protect DNNs, but conventional ADs are trained indiscriminately on benign data from all patients, overlooking physiological differences that introduce noise, degrade
arXiv:2605.12015v3 Announce Type: replace Abstract: Reusable skills are becoming a common interface for extending large language model agents, packaging procedural guidance with access to files, tools, memory, and execution environments. However, this modularity introduces attack surfaces that are largely missed by existing safety evaluations: even when the user request is benign, unsafe influence
arXiv:2605.18414v3 Announce Type: replace Abstract: Large language models increasingly operate as autonomous agents that select and invoke tools from large registries. We identify a critical gap: when unauthorized tools are visible in an agent's context, models select them in 48-68% of adversarial scenarios, even when explicitly instructed not to. Role escalation attacks (e.g., "I'm the CFO, overr
arXiv:2605.28112v2 Announce Type: replace Abstract: Federated Retrieval-Augmented Generation (FedRAG) is attractive for privacy-sensitive applications because full local corpora remain on clients. As a result, routing must rely on client-provided semantic profiles, creating a new opportunity for manipulation. We introduce Routing Hijacking, a routing-stage attack in which a malicious client forges
arXiv:2606.17092v2 Announce Type: replace Abstract: Agentic systems are increasingly integrated with geographic information systems (GIS), where multi-agent coordination enables complex conversational and spatial analysis but introduces security risks. This work presents a security-oriented framework for risk identification, evaluation, and mitigation in a multi-agent GIS system while maintaining
arXiv:2608.14800v2 Announce Type: replace Abstract: With the growth of digital document exchange, protecting image integrity against attacks such as Vector Quantization (VQ) and collage has become critical. Existing methods are vulnerable to these attacks and limited to fixed image dimensions. This paper presents a novel, dimension-agnostic, fragile watermarking algorithm that enhances security an
arXiv:2603.12753v2 Announce Type: replace-cross Abstract: Absolute anonymization, conceived as an irreversible transformation preventing re-identification and sensitive value disclosure, has proven to be a broken promise. Modern data protection must therefore shift toward a privacy-utility trade-off grounded in risk mitigation. Differential Privacy (DP) offers a rigorous mathematical framework for
arXiv:2608.27575v1 Announce Type: new Abstract: Bitcoin inverse options, traded on the Deribit exchange and settled in the underlying cryptocurrency rather than in fiat currency, combine extreme and genuinely rough volatility dynamics with a non-linear, currency-dependent payoff structure. This paper develops and empirically validates a pricing and calibration framework for these instruments based
arXiv:2608.27734v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used to discover trading strategies, and much of the resulting literature shares a methodological weakness: many candidate strategies are generated, the best is reported, and neither look-ahead bias nor the intensity of the search behind the reported result is corrected for. We present a strategy-discover
arXiv:2608.27980v1 Announce Type: new Abstract: South Korea has the world's lowest fertility and an intense education race. Families devote nine percent of lifetime income to education, mostly to private tutoring with near-zero measured returns. I show that competition for coveted careers generates an assignment externality: when all families spend more, the admission bar rises and children become
arXiv:2608.28397v1 Announce Type: new Abstract: We develop a market-informed valuation framework for guaranteed minimum maturity benefit (GMMB) riders with rational surrender under the Heston stochastic-local volatility (SLV) model. The guarantee is written on the fee-deducted account value and is considered both in its terminal-only form and in the presence of early surrender rights. The Heston S
arXiv:2608.28399v1 Announce Type: cross Abstract: In financial markets, a sequential policy that reacts systematically to price movements may become predictable to other market participants. This paper studies whether large language model (LLM) agents exhibit such directional structure through RetailAgent, an experimental framework in which an LLM observes anonymized intraday equity price historie
📰 24-Jähriger stirbt bei Motorradunfall – und wird erst Stunden später gefunden (€) Eine Fußgängerin hat am Sonntag ein Motorrad in einer Böschung bei Goldbach entdeckt. Die Polizei fand den toten Fahrer. Wann der Unfall passierte, ist unklar. Die Verkehrspolizei bittet um Hinweise. https://www. swp.de/lokales/crailsheim/zwis chen-crailsheim-und-waldtann-24-jaehriger-stirbt-bei-motorradunfall-und
The creepy comeback of smart glasses Smart glasses and facial recognition were already a reality ten years ago. Yet they were frowned upon both politically and socially. That has changed – which is why those gadgets must be banned. Just a few years ago, for many people the idea was unnerving of putting on a pair of glasses that provided additional information about objects and people. Google Glass
There's a genre of essay circulating in Taiwan-watcher spaces arguing that the responsible, unsentimental move is to negotiate unification terms while Taipei still holds leverage. I think its diagnosis is largely correct. I wrote about why the prescription still fails, on its own accounting. The argument's home turf is cold arithmetic: relative GDP, defence budgets, the US turned extractive. Fair
The bond market is growing more interesting lately — and that's concerning. Why it matters: The $160 trillion global bond market is like the plumbing in your house. You don't think about it until it stops working and you've got a nasty situation on your hands. The big picture: Long-term government bond yields for the U.S. and other G7 countries have been climbing — hovering at levels last seen in
Check out today's Metacurity for the most critical infosec developments you might have missed over the weekend, including --Infostealers hijack Claude accounts and drain users’ usage, --FulcrumSec claims Manchester airports breaches, --Rhysida auctions Berlin government data, --Bank of England warns of frontier AI risks, --DIA insider admits attempted espionage, --IPTV pirate gets six years in pri
Hi HN, we’re Brandon and Kingston, the founders of Hebbian Robotics. We built HFlow ( https://github.com/Hebbian-Robotics/hflow ), an SDK that turns multimodal recordings from robots and human operators into standardized, quality-checked episodes and queryable dataset manifests. A recording can contain synchronized video, joint states, actions, timestamps, and metadata, and HFlow processes those s
🚨🇹🇭 Thailand Cyber Police admin access allegedly advertised on a cybercrime forum ⠀ Thailand Cyber Police, a Thai law enforcement organization focused on cybercrime investigations, is allegedly affected by a security compromise after a threat actor advertised administrator-level access to an internal web panel. ⠀ The advertised access allegedly includes: ⠀ • Live access to cybercrime reports •
Bill Grueskin took time off from his work as visiting dean of Columbia Journalism School to work four months as an unpaid visiting editor at the Salt Lake Tribune. https://www. cjr.org/first_person/the-69-ye ar-old-rookie-grueskin-salt-lake-tribune-junior-editor-utah-local-newsroom-lessons.php > For all its financial and editorial success, for all its efforts to embrace the digital world, the _Sal
Students returning to Russian schools this week will discover a change to their civics courses that puts military training at the heart of the lessons, as Bloomberg [reports](https://www.bloomberg.com/news/articles/2026-08-29/russia-is-teaching-its-children-to-fight-the-wars-of-the-future) (paywalled). The mandatory classes for teenagers include learning about the main types of armaments including
Znowu odkładasz aktualizację telefonu na później? Większość z nas tak robi, ale w dzisiejszym świecie to ogromny błąd. W najnowszym wpisie tłumaczę, dlaczego ignorowanie tych powiadomień to jak zatrzaśnięcie drzwi przed darmowym ślusarzem, który chce uratować Twoje dane. Przy okazji gigantyczne dzięki! Poprzedni wpis z serii o cyfrowej higienie dotyczący aliasów email przekroczył właśnie 300 odczy
Recent AI engineering discussions centered on a production problem: large language model systems can look well-formed, cheaper, or improved while still failing on correctness, reliability, or total cost. Towards Data Science said an LLM can return valid structured JSON and still be wrong when the underlying data is messy or incomplete [towardsdatascience.com#1][google.com#1]. Reddit posts in machi
Years ago, I hand-wavingly wrote some words about how a signed WDAC policy can prevent BYOD attacks . I did not feel great with the knowledge that this isn't a thing that mere mortals can easily do, but so be it. Also, I've lived in the "Do as I say, not as I do" world for about as long as I've been in this industry. Well, I'm happy to say that times have changed. Moderately technical users that c
Centern För de företag som köper in och använder AI-lösningar ser vi det annorlunda. Vi vill att fler svenska företag ska använda AI för att bli mer effektiva och konkurrenskraftiga. Att då införa särskilda skatter eller krångliga regler skulle motverka det målet. - Uppmuntra användning. Vårt fokus ligger på att få fler företag att se och använda de möjligheter som AI ger. Det kan stärka hela Sver
Enterprise AI security vendors and researchers reported new controls, partnerships and vulnerability findings focused on autonomous agents that can act on company data and production systems. Broadcom introduced a secure, AI-ready data foundation in VMware Tanzu Platform at VMware Explore, saying the updated platform combines private-cloud data processing, sandboxed agents, isolated credentials, c
The Chief Electricity Officer used to be a real executive job: someone senior, responsible for running your own dynamo, literally keeping the lights on. The role lasted about twenty years. Not because electricity stopped mattering, but because the grid arrived and the job of minding the lights turned into paying a bill. I keep wondering if the CISO is headed the same direction, my essay runs it th
# landtagswahl # sachsen_anhalt In 6 Tagen wissen wir dann, was dieses inhaltliche Stellen und in Talkshows einladen gebracht haben wird. Es wird total spannend. 🤓👌 In ein paar Tagen danach wissen wir dann - noch mehr als jetzt schon - dass 1) das Verbotsverfahren gegen extreme Parteien ganz heftig überfällig ist, 2)unser aktueller Bundeskanzler keine Fehlerkultur hat 😘 und 3) so gar nicht mit
RE: https:// hachyderm.io/@gollyhatch/11718 0960978424676 Second round of the habanero countdown! Today I made "Swamp Soup", which I think is a dish from Louisiana and got its name because it looks like stuff floating around in the bayous. I'm not gonna link to a recipe because I changed pretty much everything from the one I used originally when I made this the first time. All recipes I found use
🚨 Robert Finale Editions WordPress dataset and source code allegedly leaked on a cybercrime forum ⠀ Robert Finale Editions, a digital gallery and corporate catalog for contemporary romantic impressionist artist Robert Finale, is allegedly affected by a data breach after a threat actor published what they claim is material stolen from the company’s WordPress website. ⠀ The advertised material incl
We, the U.S. Fish and Wildlife Service (FWS or Service), open or expand hunting opportunities on 111 field stations, including 107 units of the National Wildlife Refuges System (Refuge System or NWRS) and 4 units of the National Fish Hatchery System (Hatchery System or NFHS). This includes opening hunting or sport fishing opportunities for the first time on 14 National Wildlife Refuges (NWR) and 3
The thing that I find absolutely striking about this story is that the cops entered a pasture both without notification to the property owner and without any apparent skill or training in handling livestock. Meaning that they had no support from animal control. No training in animal control. No awareness of animal behaviors. Nothing. Apparently not even any common sense. They "felt threatened" by
It's a tough time to be a friend of Israel. Some of my Followers--especially more recent ones--may find it astounding to see me characterize myself that way. I regularly post & repost scathing indictiments of # IsraelAtWarCrimes & its ongoing # genocide in Palestine (with accompanying # EthnicCleansing in other neighboring nations) under the guise of retaliation for Palestinians' most effective ar
Soooo what self-hosted password managers are people using these days? I'm keen to self host, for centralized, cross-platform (linux, mac, maybe windows, mobile ios, maybe mobile android) support. Browser plugin for password filling a big plus, TOTP a plus. Chatbot-coded a huge minus. My fallback here is gpg-encrypted files that are a pain to use with mobile and cross-platform clients. I have been
The study, led by Ryan Thombs of Penn State University and published in PLOS Climate on July 2, drew on publicly available data from the DOD spanning 1975 to 2022, linking reductions in military spending to significant decreases in energy consumption and greenhouse gas emissions. Researchers found that a decrease in expenditures has a larger effect on slashing energy consumption than a rise in exp
arXiv:2608.28794v1 Announce Type: new Abstract: Our work evaluates the effectiveness of automated methods for solving CAPTCHA challenges commonly encountered in darknet environments. These CAPTCHAs are typically designed to operate without JavaScript, resulting in distinct characteristics compared to mainstream CAPTCHA systems. Our study considers three representative challenge types: open-circle
arXiv:2608.28882v1 Announce Type: new Abstract: Detecting sophisticated SQL Injection (SQLi) attacks remains among the most critical challenges in web applications security. This research study has resulted in an optimised hybrid BERT-GNN pipeline with improved detection accuracy and robustness while reducing false-positive and false-negative rates. SQL queries are tokenised and encoded into conte
arXiv:2608.28921v1 Announce Type: new Abstract: Behavioral biometric authentication (BBA) systems use deep learning models to transform biometric signals, such as eye movements, voice, keystroke/touchstroke dynamics, and gait, into identity embeddings for user authentication. While designed to encode identity, these embeddings may inadvertently reveal sensitive demographic attributes, including ge
arXiv:2608.28929v1 Announce Type: new Abstract: Large-scale diffusion models have fueled numerous profitable downstream applications for AI-related businesses, including visual editing and content creation. Meanwhile, due to the huge amount of resource consumption (e.g., computation and high-quality data) during training, such diffusion models are deemed valuable intellectual property (IP) for tec
arXiv:2608.28992v1 Announce Type: new Abstract: Federated learning (FL) enables collaborative training without directly sharing raw data, but remains vulnerable to malicious clients. Voting-based FL improves robustness by partitioning clients into groups, training one model per group, and aggregating predictions by plurality voting. However, under class-disjoint non-IID data, distribution-obliviou
arXiv:2608.29084v1 Announce Type: new Abstract: Face authentication is widely deployed in security-sensitive applications, while increasingly realistic 3D spoofing attacks pose growing threats. High-fidelity 3D masks can reproduce facial appearance and geometry but cannot replicate the intrinsic physical responses of living tissue, which can be actively probed by wireless signals. However, the res
arXiv:2608.29111v1 Announce Type: new Abstract: Applications such as personalized assistance and proprietary document analysis require large language models (LLMs) to generate outputs from private data. Yet powerful LLMs typically cannot be deployed on the resource-constrained devices where private data resides, and uploading private data to cloud-hosted LLMs exposes sensitive information. Recent
arXiv:2608.29134v1 Announce Type: new Abstract: Security-token standards expose privileged transfer, freezing, recovery, and compliance mechanisms, but a mechanism does not by itself identify the legal effect being executed or the evidence and reversal obligations attached to it. We formalize in Isabelle/HOL a reference execution semantics for the six regulatory-action meanings proposed in ERC-831
arXiv:2608.29136v1 Announce Type: new Abstract: We ask whether a model protects a user in the same way when that user speaks rather than types. Using a single distress vignette---a physical injury of unstated severity following an interpersonal conflict---we present four frontier models with matched inputs across voice, text, and raw API deployment conditions (n=30 per cell) and code each response
arXiv:2608.29151v1 Announce Type: new Abstract: Large Language Model (LLM) watermarks provide a mechanism for text provenance, enabling model owners to identify machine-generated content and attribute it to a specific watermarked model. However, current LLM watermarking approaches predominantly rely on inference-time sampler methods and focus their analysis on dense models. Inference-time methods
arXiv:2608.29191v1 Announce Type: new Abstract: However, no known work considered the functionality requirement in its most realistic setting. We propose a new security definition of BAEKS in the multi-user (with adaptive corruptions) and multi-challenge (both in terms of ciphertext and trapdoor in an interleaved manner) settings. We also study the question of the unforgeability of BAEKS. In fact,
arXiv:2608.29212v1 Announce Type: new Abstract: Existing video watermarking systems are symmetric: the party that can verify a mark holds the extractor weights or generator secret and can therefore also embed one. Benchmarks confirm the consequence, reporting that white-box forgery defeats all evaluated methods. We present a training-free video watermark that removes the shared secret. The signer
arXiv:2608.29381v1 Announce Type: new Abstract: AI agents are moving toward persistent, stateful execution across various applications, accumulating execution state and external effects that are costly to reconstruct after failures. Checkpoint and rollback (C/R) are becoming essential for recovery, yet their security implications remain largely unexplored. Correct rollback does not imply secure re
arXiv:2608.29463v1 Announce Type: new Abstract: A benchmark score is a joint property of the model, the evaluation harness, the elicitation budget, the sampled population, and contamination status. Leaderboards publish the model and the score, so capability and leakage stay observationally equivalent. Existing taxonomies classify contamination for automated detection, not the question a reporter f
arXiv:2608.29568v1 Announce Type: new Abstract: Different attack methods follow different search trajectories, they succeed on different subsets of samples, whereas existing hard-label black-box text attacks mainly focus on improving individual attackers or manually combining them. We present {\OURS}, a method for optimizing attacker sequences in hard-label black-box text attacks. {\OURS} first pe
arXiv:2608.29745v1 Announce Type: new Abstract: LLMs are widely deployed through cloud-hosted inference services, where Just-in-Time (JIT) compilation is used to reduce recurring framework and GPU-launch overhead. JIT serving introduces a host-side control plane that selects compiled artifacts and orchestrates their execution on the GPU. Meanwhile, the shared cloud setting has motivated a growing
arXiv:2608.29758v1 Announce Type: new Abstract: The increasing number of vulnerabilities in operating systems, together with sophisticated kernel-level threats (e.g., rootkits), has weakened the effectiveness of traditional in-kernel protection mechanisms. Since these defenses operate at the same privilege level as the kernel, they share the same attack surface and can be bypassed once the kernel
arXiv:2608.29773v1 Announce Type: new Abstract: Hardware Security Module (HSM) logs capture security-critical behavior, but anomalies emerge from relationships across event sequences, keys, object states, sessions, and temporal patterns rather than isolated events. Existing methods separate detection from HSM-specific evidence validation and reporting. In this paper, we present HSMLog, a two-stage
arXiv:2608.29942v1 Announce Type: new Abstract: The key limitation of current state-of-the-art influence-based guardrails is that they do not reliably distinguish a legitimate, user-authorized action from a malicious, unauthorized action when both rely on external tool information. This ambiguity can cause benign actions to trigger unnecessary verification and intervention, reducing utility and ad
arXiv:2608.29977v1 Announce Type: new Abstract: The Learning With Errors (LWE) problem, introduced by Regev in 2005, is central to modern cryptography and post-quantum security. The algorithms to solve the search version of the problem, Search-LWE, can be broadly categorised into algebraic, combinatorial and lattice-based. In this work we propose a new algebraic algorithm for the Search-LWE proble
arXiv:2608.30038v1 Announce Type: new Abstract: System-level mobile agents are evolving from fixed scripts into adaptive systems that continuously observe interfaces, reason, and adjust their actions, allowing automated attacks to navigate dynamic UIs and complete complex tasks. Existing applications detect automation using touch trajectories, action timing, and the physical coupling between touch
arXiv:2608.30041v1 Announce Type: new Abstract: Large language model agents place outputs from external skills into their execution context, allowing attacker-controlled data to influence later privileged actions. Existing defenses mainly classify untrusted content or authorize proposed operations. They do not directly address how an agent's future authority should change once untrusted data enter
arXiv:2608.30083v1 Announce Type: new Abstract: Multi-agent AI platforms move quickly from staging to production, but the way agents establish trust remains rudimentary: an agent either transmits raw data to a peer or accepts that peer's natural-language self-report that a value complies with policy. The first over-shares; the second is unverifiable and is exactly the channel prompt injection atta
arXiv:2608.30141v1 Announce Type: new Abstract: Preference optimization is widely used to align large language models with human preferences, but preference-data composition may also influence privacy-relevant memorization. We examine whether adding synthetic privacy-preference pairs to Direct Preference Optimization (DPO) is associated with lower canary-based memorization signals without modifyin
arXiv:2608.30177v1 Announce Type: new Abstract: Long-term memory is becoming a core capability of LLM agents, enabling personalization and long-horizon interaction. However, memory mechanisms that retain, transform, or expose more information can affect both benign utility and susceptibility to memory poisoning. Existing evaluations typically measure memory utility or attack risk in isolation unde
arXiv:2608.30207v1 Announce Type: new Abstract: Computer use agents (CUAs) are vision-language models that perceive a screen and act on a real operating system through mouse, keyboard, and terminal, and they are increasingly deployed to automate everyday digital tasks. Because they can be exposed to untrusted content while operating, they are vulnerable to indirect prompt injection (IPI), in which
arXiv:2608.30288v1 Announce Type: new Abstract: LLM agents commonly use knowledge-based tools and access their underlying files, databases, and search indexes through tool invocation. This integration improves agents' ability to provide domain-specific services but also introduces the risk of tool-mediated knowledge extraction: source content exposed to an agent for legitimate responses may be pro
arXiv:2608.30387v1 Announce Type: new Abstract: Multi-agent applications delegate work across independently operated deployers. After an incident, a verifier must answer two questions: which deployer released the reported bytes, and whether each cross-deployer edge was authorized. Credentials establish who may act, but need not bind them to later output bytes or prove both deployers authorized a d
arXiv:2608.30403v1 Announce Type: new Abstract: Backdoor attacks pose a serious threat to large language models (LLMs), but existing defenses remain fragmented, failing to pro?vide unified defense against both dirty-label and clean-label attacks. To investigate why such fragmentation arises, we present the first systematic feature-level mechanistic analysis of LLM backdoors using sparse autoencode
arXiv:2608.30441v1 Announce Type: new Abstract: Recently, large language model (LLM) agents, such as Codex, Claude Code, and OpenClaw, have become capable of planning and executing long-horizon tasks through repeated tool calls. This capability also creates new opportunities for prompt injection. Existing attacks either place the malicious objective in one explicit instruction, making it easy to d
arXiv:2608.30648v1 Announce Type: new Abstract: Zero-knowledge domain-specific language (ZK DSL) toolchains compile programs into constraint systems and generate witnesses for cryptographic proofs. Bugs in these toolchains can leave the enforced constraints weaker than the source-program semantics, admitting proofs for invalid executions. Such soundness bugs may remain invisible to valid-execution
arXiv:2608.30686v1 Announce Type: new Abstract: Coding agents are increasingly used for software engineering tasks, including bootstrapping projects from third-party repositories whose integrity cannot be assumed. Prior work on repository poisoning largely focuses on attacker-controlled injection and disguise, but developers also shape risk through everyday invocation choices: what task to delegat
arXiv:2608.30969v1 Announce Type: new Abstract: Systems for speaker anonymization obfuscate the speaker of an utterance, while maintaining its original semantic contents and prosody. Recent solutions for speaker anonymization rely on learned representations that disentangle an utterance into semantic contents and speaker properties. To anonymize an utterance, these systems replace the speaker prop
arXiv:2608.28732v1 Announce Type: cross Abstract: Ensuring the security of complex systems involves the strategic allocation of defensive resources to prevent various types of attacks from succeeding. A defender often has multiple types of defensive assets at its disposal, where it must decide how to optimally deploy their heterogeneous capabilities across different attack types. In this paper, we
arXiv:2608.28843v1 Announce Type: cross Abstract: We show that smooth two-layer feed-forward networks (FFNs) expose an additional structural model extraction channel under a chosen-input raw-output oracle at the FFN branch; consider transformer FFN branches with GELU or SiLU activations under chosen-input raw-output access, without access to parameters, gradients, or internal activations; exploit
arXiv:2608.29460v1 Announce Type: cross Abstract: When coding agents encounter defective test infrastructure they may reward-hack: hardcoding outputs or editing test files to pass tests they cannot legitimately satisfy, a pattern that has now appeared outside benchmarks, in a coordinated multi-agent intrusion of a major AI platform's production infrastructure. The same capability that lets an agen
arXiv:2608.29510v1 Announce Type: cross Abstract: Aerial object detection is increasingly deployed in real-world applications, but models remain vulnerable to physical, universal adversarial patches that cause them to miss objects. Furthermore, defenders face the practical constraint of training data scarcity: aerial imagery is costly to collect and label, so a deployment site typically yields hun
arXiv:2608.29851v1 Announce Type: cross Abstract: The impact of Python applications has been evidenced by their widespread presence in some of the most impactful software domains, such as machine learning frameworks and scientific computing platforms. These applications often integrate native code components written in a lower-level programming language like C. This multilingual construction bring
arXiv:2608.30329v1 Announce Type: cross Abstract: Speech enhancement models are widely deployed as frontend modules in real-time speech services, yet their vulnerability to backdoor attacks remains unexplored. Existing backdoor methods are confined to classification tasks and rely on active trigger injection, an assumption incompatible with the passive processing nature of speech enhancement model
arXiv:2608.30473v1 Announce Type: cross Abstract: Differential privacy has become the standard for generating privacy-protected data releases. However, differential privacy does not translate intuitively to disclosure risk. In particular, it remains unclear how much an adversary's belief about an individual's inclusion in a dataset can change after observing a protected release. To address this qu
arXiv:2608.30519v1 Announce Type: cross Abstract: AI agents can select tools, counterparties, and transaction parameters, yet inference should not itself confer authority to execute a financial action. This study develops and evaluates Authority-Inference Separation (AIS), an intent-centered architecture for bounded agentic finance. AIS treats a financial action intent as the control object: a mac
arXiv:2608.31112v1 Announce Type: cross Abstract: We obtain a certified randomness protocol in the quantum random oracle model. The protocol is non-interactive and publicly verifiable with a classical verifier, and is based on Yamakawa and Zhandry's proof of quantumness [JACM'24]. We prove unconditional security of this protocol against adversaries making subexponentially-many adaptive quantum que
arXiv:2608.31142v1 Announce Type: cross Abstract: The 2025--2026 AI market has seen a wave of stealth releases: frontier models launched anonymously on developer platforms under codenames. For their users, identity determines data-handling terms, supply-chain risk, and capability expectations. No validated methodology exists for black-box identity verification of anonymous models: practitioner che
arXiv:2608.31150v1 Announce Type: cross Abstract: We rethink the definition of privacy in multi-server, graph-replicated private information retrieval (PIR) systems, by introducing a novel setting where the user's privacy is governed by the servers' storage structure. In classical graph-replicated PIR, the user retrieves a single message stored at the servers, while hiding the message index from e
arXiv:2502.10673v2 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) has become an effective method for enhancing large language models (LLMs) with up-to-date knowledge. However, it may pose a risk of copyright infringement, as IP datasets may be incorporated into the knowledge database by malicious Retrieval-Augmented LLMs (RA-LLMs) without authorization. To protect the rights
arXiv:2503.11841v2 Announce Type: replace Abstract: Machine Learning (ML) malware detectors rely heavily on crowd-sourced AntiVirus (AV) labels, with platforms like VirusTotal serving as trusted sources of malware annotations. But what if attackers could manipulate these labels to classify benign software as malicious? We introduce label spoofing attacks, a new threat that contaminates crowd-sourc
arXiv:2503.12896v2 Announce Type: replace Abstract: Recent studies improve on-device language model (LM) inference through end-cloud collaboration, where the end device retrieves useful information from cloud databases to enhance local processing, known as Retrieval-Augmented Generation (RAG). Typically, to retrieve information from the cloud while safeguarding privacy, the end device transforms o
arXiv:2506.11635v2 Announce Type: replace Abstract: Credit card fraud mitigation plays a significant role in modern society. While fraud detection systems are essential, they often struggle to keep pace with the constantly evolving fraud techniques. As a result, fraud investigation is an important complementary process required for continuously improving detection models, identifying emerging frau
arXiv:2507.05512v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used for code generation, making reliable identification of machine-generated code important for attribution, tracking, and misuse detection. Existing code watermarking methods are dominated by N-gram-based schemes, yet their robustness has mostly been evaluated only against simple edits or optimizati
arXiv:2508.21457v4 Announce Type: replace Abstract: The rapid advancement of Large Language Models (LLMs), with their growing abuse in phishing, has enabled phishing content, including deceptive pretexts and other persuasive elements, to be generated at a scale difficult to achieve manually. This growing misuse of LLMs in phishing raises questions about potential LLM-driven changes in phishing cha
arXiv:2509.03985v3 Announce Type: replace Abstract: Jailbreak attacks bypass the safety alignment of large language models (LLMs) to elicit harmful outputs, yet the vast parameter space makes diagnosing the underlying failure mechanisms extremely challenging. We present NeuroBreak, a visual analytics system that helps experts progressively unpack jailbreak mechanisms from layer-level semantics dow
arXiv:2510.01359v3 Announce Type: replace Abstract: Code-capable large language model (LLM) agents are embedded in software engineering workflows where they can read, write, and execute code, raising "jailbreak" stakes beyond text-only settings. Prior evaluations emphasize refusal or harmful-text detection, leaving open whether agents compile and run malicious programs. We present JAWS-Bench (Jail
arXiv:2511.14088v3 Announce Type: replace Abstract: Run-time integrity enforcement in real-time systems presents a fundamental conflict with availability. Existing approaches in real-time systems primarily focus on minimizing the execution-time overhead of monitoring. After a violation is detected, prior works face a trade-off: (1) prioritize availability and allow a compromised system to continue
arXiv:2512.08289v4 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) systems enhance LLMs with external knowledge but introduce a critical attack surface: corpus poisoning. While recent studies have demonstrated the potential of such attacks, they typically rely on impractical assumptions, such as white-box access or known user queries, thereby underestimating the difficulty of
arXiv:2602.11793v2 Announce Type: replace Abstract: Watermarking has emerged as a crucial technique for detecting and attributing content generated by large language models. While recent advancements have utilized watermark ensembles to enhance robustness, prevailing methods typically prioritize maximizing the strength of the watermark at every individual layer. In this work, we identify a critica
arXiv:2603.22341v2 Announce Type: replace Abstract: While prior red-teaming efforts have focused on eliciting harmful text outputs from large language models (LLMs), such approaches fail to capture agent-specific vulnerabilities that emerge through multi-step tool execution, particularly in rapidly growing ecosystems such as the Model Context Protocol (MCP). To address this gap, we propose a traje
arXiv:2604.12431v3 Announce Type: replace Abstract: Organisations increasingly outsource privacy-sensitive data transformations to cloud providers, yet no practical mechanism lets the data owner verify that the contracted algorithm was faithfully executed. VeriX-Anon is a multi-layered verification framework for outsourced Target-Driven k-anonymization combining three orthogonal mechanisms: determ
arXiv:2604.14330v3 Announce Type: replace Abstract: "Cryptography turns a security problem into a key management problem." Despite decades of research effort towards usable key management, it remains unclear whether key management issues are inherent to every cryptographic system or merely artifacts of specific designs. To investigate, this paper presents a user study of mutual TLS (mTLS) usabilit
arXiv:2604.27426v2 Announce Type: replace Abstract: Local fine-tuning datasets routinely contain sensitive secrets such as API keys, personal identifiers, and financial records. Although "local offline fine-tuning" is often viewed as a privacy boundary, we reveal that compromised model code is sufficient to steal them. Current passive pretrained-weight poisoning attacks, while effective for natura
arXiv:2605.22122v2 Announce Type: replace Abstract: Collaborative perception (CP) enables connected and autonomous vehicles to share sensor data and jointly reason about their environment. To defend against adversaries that fabricate or manipulate shared data, existing systems employ cross-vehicle inconsistency detection and trust estimation, penalizing vehicles whose observations conflict with th
arXiv:2605.30123v3 Announce Type: replace Abstract: Homomorphic encryption (HE) enables privacy-preserving aggregation in federated learning (FL) by allowing the server to operate on encrypted data without decryption. Existing HE-over-the-air (OTA) methods mainly rely on single-key HE schemes and require channel estimation or pre-equalization to compensate for wireless fading. However, single-key
arXiv:2606.01442v2 Announce Type: replace Abstract: Spiking neural networks (SNNs) are increasingly studied for network intrusion detection, but comparative evidence on how neuron models and spike encodings affect performance remains limited. Evaluation choices can influence results when preprocessing, capture structure, or scenario information crosses the train--test boundary. We evaluate nine sn
arXiv:2606.21210v3 Announce Type: replace Abstract: Acoustic side-channel attacks (ASCA) on keyboards have gained increasing attention, yet impact of speech representation learning models in ASCA remains unexplored. Addressing this, we introduce KEYAC, a dataset designed to analyze representation generalization for ASCA under both standard and VoIP codec settings. On KEYAC, we evaluate six represe
arXiv:2608.02665v2 Announce Type: replace Abstract: A benchmark score is a measurement instrument, yet most benchmarks read each item at a single canonical surface form. We ask whether that reading is faithful: when an item's intent is held fixed and only its meaning-preserving surface form varies, does the canonical-form score estimate model behavior well, and how much of any variation is decodin
arXiv:2608.16032v2 Announce Type: replace Abstract: LLM agents are stateless and rely on external memory to carry context between steps. Because agents treat that memory as trustworthy, an adversary who can write to it can steer their behavior. The FARMA attack does this with no malicious command: it inserts fabricated entries into the agent's reasoning memory claiming a required safety step is al
arXiv:2402.03741v4 Announce Type: replace-cross Abstract: Recent advancements in multi-agent reinforcement learning (MARL) have opened up vast application prospects, such as swarm control of drones, collaborative manipulation by robotic arms, and multi-target encirclement. However, potential security threats during the MARL deployment need more attention and thorough investigation. Recent research
arXiv:2405.18540v3 Announce Type: replace-cross Abstract: Red-teaming, or identifying prompts that elicit harmful responses, is a critical step in ensuring the safe and responsible deployment of large language models (LLMs). Developing effective protection against many modes of attack prompts requires discovering diverse attacks. Automated red-teaming typically uses reinforcement learning to fine-
arXiv:2505.16567v4 Announce Type: replace-cross Abstract: Finetuning open-weight Large Language Models (LLMs) is standard practice for achieving task-specific performance improvements. Until now, finetuning has been regarded as a controlled and secure process in which training on benign datasets leads to predictable behaviors. In this paper, we demonstrate, for the first time, that an adversary ca
arXiv:2605.12729v3 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly being used in network operations (NetOps) and artificial intelligence for IT operations (AIOps) for tasks ranging from telemetry retrieval and incident diagnosis to configuration planning and bounded remediation. As these systems acquire greater access to operational tools, the central question
arXiv:2605.29224v2 Announce Type: replace-cross Abstract: AI agents augment large language models with external tools such as web retrieval, enabling grounded and up-to-date responses. However, incorporating external content into the generation pipeline can weaken the safety alignment mechanisms that govern model outputs. Prior work shows that enabling retrieval in agents increases compliance with
arXiv:2608.29025v1 Announce Type: new Abstract: Classical option-hedging methods like Black-Scholes delta assume constant, free rebalancing, which real markets don't allow. Deep hedging trains a neural network to handle these frictions directly, and prior work reports strong results. But those comparisons usually pit deep hedging against a frictionless classical baseline on simulated price data. T
arXiv:2608.29468v1 Announce Type: new Abstract: We study the quadratic tracking problem of a general stochastic target process with absolutely continuous controls, with and without terminal constraint. We derive explicit, non-asymptotic upper bounds in terms of a Besov-type modulus of the target. These bounds yield sharp explicit rates that specialize to the square-root order for semimartingale ta
arXiv:2608.29669v1 Announce Type: new Abstract: Spatial return models take the interaction matrix as given and leave feedback uninterpreted. We construct a bandwidth-free field from firms' language-model article embedding distributions using target-anchored Wasserstein barycentric reconstruction. A quadratic exposure-adjustment problem maps feedback into a peer-misalignment penalty ratio. For 52 f
this outlet adds: 29669v2, replace, asset, pricing, structure, inter, firm, infer
arXiv:2608.29786v1 Announce Type: new Abstract: This paper develops a latent-state framework for recovering borrower-level posterior beliefs in credit-risk analysis. Creditworthiness and financial fragility are represented as latent dimensions, while observed borrower scores follow a finite Gaussian mixture model and default depends on the latent profile. Borrower-specific probabilities of default
arXiv:2608.30490v1 Announce Type: new Abstract: Claims of the form "we find no evidence that X affects Y" appear throughout the applied finance literature, yet whether such a claim contains evidence of absence or absence of evidence depends entirely on its confidence interval. The term "statistically insignificant" is routinely read to mean zero economic effect. However, a more honest description
arXiv:2608.30522v1 Announce Type: new Abstract: Tariff threats can move household beliefs before policy is enacted, yet their rapidly changing language is difficult to study with conventional surveys. We build a multi-agent system that turns 300 households from the Michigan Surveys of Consumers into persistent large-language-model agents exposed to social-media information over several simulated m
arXiv:2608.30867v1 Announce Type: new Abstract: We propose a neural calibration method to construct a recombining binomial tree directly from a set of given option prices. Rather than estimating a continuous option pricing function or a local volatility surface as an intermediate object, a neural network is used to deform a benchmark lattice. This leads to a discrete pricing model which is guarant
arXiv:2608.30999v1 Announce Type: new Abstract: Market-order flow in financial markets exhibits long-range correlations. This is a widely known stylised fact of financial markets. A popular hypothesis for this stylised fact comes from the Lillo-Mike-Farmer (LMF) order-splitting theory. However, quantitative tests of this theory have historically relied on proprietary datasets with trader identifie
arXiv:2608.31041v1 Announce Type: new Abstract: Quantitative trading is moving from isolated predictive models toward agentic workflows that combine reasoning, tool use, memory, and feedback. This survey reviews agentic quantitative trading across five stages: factor mining, signal discovery, portfolio construction, order execution, and risk management. We further examine agentic quant trading sys
arXiv:2405.11392v3 Announce Type: replace Abstract: We propose a deep learning algorithm for high dimensional optimal stopping problems. Our method is inspired by the penalty method for solving free boundary PDEs. Within our approach, the penalized PDE is approximated using the Deep BSDE framework proposed by \cite{weinan2017deep}, which leads us to coin the term "Deep Penalty Method (DPM)" to ref
arXiv:2410.23447v2 Announce Type: replace Abstract: Firm characteristics are commonly represented as fixed vectors, even though evidence about firms' operations arrives as heterogeneous collections of articles reports. We study how distances between distributions of firm characteristics restrict systematic covariance. For given latent exposure laws, quadratic Wasserstein geometry yields sharp cova
arXiv:2603.10807v2 Announce Type: replace Abstract: Existing LLM safety evaluations rely on binary attack-success rates and domain-agnostic taxonomies, leaving regulated Banking, Financial Services, and Insurance (BFSI) deployments exposed to failures elicited through legally or professionally plausible framing. We introduce RAHS (Risk-Adjusted Harm Score), a risk-sensitive metric jointly capturin
arXiv:2604.27700v2 Announce Type: replace Abstract: The rapid growth of weather-dependent renewable generation increases price volatility and imbalance penalty risk in power markets, creating the need for advanced quantitative trading strategies. We develop a data-driven continuous-time stochastic optimal control framework for intraday electricity trading using stochastic differential equations wi
arXiv:2607.13844v3 Announce Type: replace Abstract: Journals cannot referee every paper they receive. When submissions outrun the time reviewers can give, an editor must decide what to do with the papers no one reads: reject them, or check them with a faster but less accurate tool, of the kind artificial intelligence now supplies. A paper the tool passes carries the same public stamp as one an exp
arXiv:2603.14557v2 Announce Type: replace-cross Abstract: Banks must optimize risky investments, dividend payouts, and capital structure under tight Basel III solvency and liquidity constraints, while costly equity issuance serves as a distress-recovery tool. We formulate this as a stochastic control problem that reduces the high-dimensional balance-sheet dynamics to a tractable one-dimensional pr
arXiv:2605.27887v4 Announce Type: replace-cross Abstract: Large language models (LLMs) have shown strong performance across diverse financial tasks, yet portfolio management (PM) remains poorly benchmarked. Existing benchmarks exhibit two gaps: they are often equity-only and ignore cross-asset correlations; they fail to evaluate the complete PM decision pipeline. We introduce PortBench, a benchmar
arXiv:2606.30085v2 Announce Type: replace-cross Abstract: Large-language models have proven to be remarkable if inconsistent parrots of public attitudes and opinions. The extent to which LLMs are able to produce reasonable approximations of cultural taste remains an open empirical question that becomes more urgent by the day, with market research companies already offering provisional 'synthetic'
🚨 Dropbox Hack / Data Breach: Lenovo ID Flaw Enabled Account Takeover A serious Dropbox security incident highlights a dangerous weakness in federated identity. Attackers allegedly registered Lenovo IDs using victims’ email addresses, then abused Dropbox SSO / OIDC federation** to authenticate against existing Dropbox accounts. The disturbing part: • No Dropbox password was required • Victims did
I hear a lot from over 30s that theyre sort of losing interest in big titles because theyre just too much of a time sink. they have about 30m to game a day and most of that is taken up by a single level Ive been telling them yeah thats why i got an FPGA because I can emulate old games where you hit play and you can play galaga for 30m. Its a reason I'm so into like Atari and Mame, because arcade g
To deliver web platform advancements, performance improvements, and bug fixes faster, Chrome's release cycle has moved to every two weeks starting this month. Beginning with Chrome 153 scheduled for release this month, the Beta and Stable channels will ship every two weeks instead of four. This applies to Desktop, Android, and iOS. Smaller-scoped releases not only get features into your hands fast
Organisational Dysfunction of the Day Deskilled by AI Context: The calculator took mental arithmetic. GPS took a sense of direction that used to build itself through years of navigating without help. Neither loss announced itself; the skill just stopped being exercised, then stopped being there. The organisation is rolling out AI coding assistants. Mandatory adoption, tracked usage, productivity m
With the end of physical PlayStation games looming, the need for clear digital ownership rights is more pressing than ever. For its part, Sony believes your rights are very clear: you do not own your digital games, and it's unlikely you're stupid enough to believe otherwise. At least, that's roughly the argument the company's lawyers are presenting in response to a proposed class-action lawsuit. S
We are scanning & reporting daily on vulnerable Microsoft Exchange CVE-2026-62911 (Authentication Bypass by Capture-replay) instances in our Vulnerable Exchange reporting: https://www. shadowserver.org/what-we-do/ne twork-reporting/vulnerable-exchange-server-report/ At least 21899 IPs seen unpatched 2026-08-31, top US (6.2K) & Germany (5.1K) Dashboard World Map view stats: https:// dashboard.shado
Turkish President Erdogan: First and foremost, the Shanghai Cooperation Organization has a structure that aligns with our multidimensional foreign policy centered on regional ownership. As Türkiye, we are taking responsibility for resolving the conflicts around us through dialogue and diplomacy, and we continue to work for peace, prosperity, and stability in our region and throughout the world. We
The US Postal Service effort to develop a sweeping new system for verifying mail ballots has invited a host of IT concerns among agency officials, with some describing the process internally as a “sh*t show,” according to a new whistleblower report released Tuesday by a Democratic senator. The unnamed whistleblower is alleging that some of the procedures USPS is planning – under mandates put forwa
@ andrewnez read that AROMA+ paper you linked to in your "the week in package management". There's a bit of their discussion I'm tempted to email them about to disagree with, specifically where they propose removing JARs from maven central because they couldn't reproduce the build any more, " find it worrisome to observe how many projects cannot be reproduced simply because the necessary resources
Oluwadamilola Ogooluwa Bamigboye, 24, and Rekeya Lionesha Lee Frazier, 23, were hit with federal charges after agents approached Bamigboye regarding an alleged student visa overstay. According to the account, the encounter quickly spiraled when Frazier got behind the wheel and drove away with one of the agents still inside the vehicle, heading toward a police station. Bamigboye, meanwhile, reporte
Exclusive: Two congressmembers respond to revelations of government snooping on journalists, non-profits and unions Two congressional Democrats asked a government watchdog Tuesday to investigate the Trump administration’s use of an obscure law to secretly collect records on journalists, non-profits and unions without judicial oversight. The request from the senator Ron Wyden of Oregon and the repr
Hi HN! My son spent a lot of time in 5th grade playing with the 2-dimensional PhET gravity/space sims[1]. His STEM class uses it, and he was playing with it at home, and has been looking for more. I found several space sims/playgrounds online, including NASA's Eyes on the Solar System[2], but nothing that seemed to offer a real progression from what he was getting out of the PhET sims, so I built
New. Broadcom has several advisories addressing a mixed bag of vulnerabilities https:// support.broadcom.com/web/ecx/s ecurity-advisory # Broadcom CISA has added several industrial vulnerabilities to the catalogue https://www. cisa.gov/news-events/ics-advis ories # CISA Google: Chrome Stable for iOS Update https:// chromereleases.googleblog.com/ # Google # Chrome Nvidia Security Bulletin: NVIDIA M
The toes! Chameleons all have weird feet with fused toes in a weird, tongs-like configuration, appearing as though they only have two toes per foot. (If you look closely, you can see that there are two claws on one "toe", and three on the other). This one has 20 well-differentiated toes. The body plan is also much more elongated than you'd expect in a chameleon - and the tail is much too slender a
Data: U.S. Census Bureau, FactSet Construction spending on AI data center "shells" soared in July, rising at an annualized rate of nearly 60% from July 2025 levels. Why it matters: It shows the AI building frenzy was gathering strength this summer, even as — or perhaps because — a bipartisan backlash grows. By the numbers: Data center construction spending jumped to an annual pace of more than $75
Dewlaps are pretty common in lizards, and they're almost always a secondary sex characteristic. For example, draco lizards and iguanas have them too - although female iguanas, being very large animals, have pretty respectable dewlaps even if they're smaller than those of males. Anoles take them to a pretty ridiculous level, where males have extremely large, conspicuous and brightly coloured ones,
In all fairness, Wikipedia's donor relations got back to me on my "When you stop union busting I'll start contributing again" replly: "Hi Otto, Thank you for your email and support of the Wikimedia Foundation and free knowledge. We appreciate you sharing your concerns with us. The Wikimedia Foundation respects employees’ rights to organize and is committed to supporting a fair election process. Th
Incertitudinea și anxietatea: ce se întâmplă în mintea noastră când nu avem răspunsuri - Mintea umană se străduiește neîncetat să găsească explicații clare pentru fiecare eveniment din viața cotidiană. Lipsa unor răspunsuri certe declanșează rapid o stare de alertă interioară, deoarece creierul interpretează necunoscutul drept o potențială amenințare la adresa stabilității. Această dinamică cognit
I made OwnTime to balance between competing priorities in my life. The idea is based on two main influences: the concept of "roles" from "The 5 Choices" (2015), and my repeated failure to effectively implement time blocking due to the necessary flexibility in my role. The app allows you to define time budgets, which are not much more than a few mutually exclusive countdown timers. The UX is essent
Bike-adjacent project vent: I recently coordinated the purchase of new bike racks (seven rail-mounted ⋂-hoops) from a manufacturer in Waunakee, WI, for use in new bike shelters at our coho. Order fulfillment took 4 weeks, ground shipping cost $550 (10x the cost of freight a decade ago) and the order was delivered a few weeks ago -- with the wrong parts. The steel rails do not match the hoops order
arXiv:2609.00052v1 Announce Type: new Abstract: Commercial LLM APIs advertise a specific foundation model, but the served backbone may be silently substituted, quantized, or wrapped, for example to save deployment costs. All existing audits decide backbone identity from the text-output channel, which is structurally fragile for agentic APIs because modern serving stacks (OpenAI, Anthropic, Gemini,
arXiv:2609.00259v1 Announce Type: new Abstract: We propose a novel approach to learning-based attack forensics called DUPIN. DUPIN performs unsupervised pre-training on an enormous amount of audit events in the form of provenance graphs. It then proceeds to a few-shot learning stage, leveraging a small number of labeled attack examples to fine-tune its detection capabilities. We pretrain DUPIN on
arXiv:2609.00267v1 Announce Type: new Abstract: Autonomous LLM agents increasingly act on a user's behalf: they hold credentials, call tools and services, and spawn sub-agents that act further on their behalf. This turns a long-standing distributed-systems question -- who is authorized to do what, on whose authority -- into an urgent and largely unsolved problem, because the component driving each
arXiv:2609.00340v1 Announce Type: new Abstract: Critical-minerals supply chains face a structural tension: regulators and buyers demand verifiable provenance, yet upstream actors are hesitant to disclose supplier identities, assay grades/yields, and prices that verification appears to require. We report a design science account of OreProof, a prototypical traceability platform addressing this veri
arXiv:2609.00390v1 Announce Type: new Abstract: Wearable EEG systems may expose sensitive information beyond their intended health function, creating substantial risks to neuroprivacy. In this work, we show that commonly used EEG features can reveal participant identity and demographic attributes in addition to supporting the intended cognitive task. Wearable EEG is increasingly being explored for
arXiv:2609.00430v1 Announce Type: new Abstract: Recent work uses large language models to generate systems code from specifications, treating the specification as the durable artifact and the implementation as disposable. Regenerating the implementation specializes it to each workload and device. However, that work lives in a forgiving setting: a component's externally visible effects, its writes
arXiv:2609.00464v1 Announce Type: new Abstract: Neuro-Symbolic (NeSy) AI has recently emerged as a novel paradigm to enable trustworthy AI, aiming at integrating sub-symbolic neural perception with grounded symbolic reasoning. The neuro-symbolic integration process that characterizes these models has been proven beneficial to achieve more transparent, explainable and efficient AI systems. Meanwhil
arXiv:2609.00502v1 Announce Type: new Abstract: Physical fault injection can turn brief hardware disturbances into security failures such as key recovery, authentication bypass, and unintended control flow. Finding effective faults is difficult because many interacting parameters create a large search space, successful settings are sparse and target-dependent, and each hardware attempt provides li
arXiv:2609.00523v1 Announce Type: new Abstract: Long-term memory can turn untrusted external content into persistent influence over an LLM agent's future decisions, creating the threat of indirect memory poisoning. A successful attack must survive a multi-stage pipeline comprising memory writing, retrieval, and utilization. Existing attacks largely rely on intra-stage optimization, optimizing indi
arXiv:2609.00595v1 Announce Type: new Abstract: Safe agents can fail together. Multi-agent LLM systems (MAS) move information, state, decisions, and authority across principal boundaries, creating failures that local checks may miss. Without an execution-level view, a multi-agent setting can easily be mistaken for evidence of a genuinely multi-agent security effect. We thus systematize MAS securit
arXiv:2609.00604v1 Announce Type: new Abstract: The growing complexity of cyber-physical attack surfaces in advanced manufacturing has made cyber threat intelligence analysis increasingly difficult. Although large language models and retrieval-augmented generation have improved CTI workflows, text-based approaches remain vulnerable to hallucinations and provide limited support for structured reaso
arXiv:2609.00609v1 Announce Type: new Abstract: Many questions about a digital circuit take the same form. A hidden object is consistent with a set of observations, and one wants to know how many remain consistent and which observation to make next. The set of surviving candidates is the version space, and its size, on a logarithmic scale, measures how much the observations have settled. This pape
arXiv:2609.00676v1 Announce Type: new Abstract: In the contemporary landscape of software devel-opment, securing sensitive data is paramount to safeguarding organizational reputation, preventing financial losses, and pro-tecting individuals from identity theft. This paper addresses the pervasive challenge of identifying and rectifying security vulnerabilities early in the development process, emph
arXiv:2609.00705v1 Announce Type: new Abstract: Prior adversarial attacks on Windows PE malware detectors target raw bytes, PE headers, or intra-function control-flow graphs, leaving the function call graph (FCG) unexplored as an attack surface. Yet the FCG structure is an important feature in graph-based malware detectors. We present Phan- tomCall, a black-box attack that perturbs the FCG of Wind
arXiv:2609.00711v1 Announce Type: new Abstract: Extended Reality (XR) provides immersive, interactive 3D experiences. To enable these experiences, the devices must track user motion so the system can respond to actions such as grabbing, looking at, or moving an object. However, motion tracking has raised privacy concerns since it records a person's motion patterns. These motion patterns have been
arXiv:2609.00790v1 Announce Type: new Abstract: Reliable refusal behavior requires Large Language Models (LLMs) to reject harmful prompts with only answering benign ones. Incorrect refusal behavior can either expose users to harmful responses or prevent users from obtaining useful answers. Training-time alignment improves refusal behavior by updating model parameters with safety data, but requires
arXiv:2609.00806v1 Announce Type: new Abstract: In 2025, scams were responsible for an estimated $442 billion in direct losses globally. In the United States, reported losses increased by nearly 400% between 2020 and 2025. Though AI in scamming is a relatively new phenomenon, its use significantly changes the economics of scams as well as the bottlenecks in scam operations. In this paper I investi
arXiv:2609.00886v1 Announce Type: new Abstract: Cyber ranges are complex environments comprising many interacting components and stakeholders with different security concerns. The Service-Oriented Cyber Range (SOR) is no exception, particularly when it comes to training scenarios targeting critical infrastructure. Security concerns are translated into security requirements, the elicitation of whic
arXiv:2609.01046v1 Announce Type: new Abstract: Production LLMs must handle inputs that attempt to override system instructions, bypass safety policies or elicit harmful responses. A common mitigation is a separate guardrail model. Existing reports, however, provide little evidence on Russian prompt injection or Russian surface obfuscation. We present HiveTraceGuard-Pro, a 0.6B generative guardrai
arXiv:2609.01077v1 Announce Type: new Abstract: Safety-critical real-time systems must satisfy multiple dependability requirements, notably time predictability and security. In such systems, tasks must complete within bounded and known execution times, typically characterised through Worst-Case Execution Time (WCET) analysis. At the same time, DRAM-based platforms are increasingly sensitive to the
arXiv:2609.01171v1 Announce Type: new Abstract: SMS spam detection systems often achieve high accuracy in controlled environments but struggle against adversarial attacks and increasingly sophisticated spam tactics in real-world deployments. In this paper, we evaluate the robustness of SMS anti-spam systems that end users actually rely on, including commercial messaging applications, third-party a
arXiv:2609.01174v1 Announce Type: new Abstract: Cyber Threat Intelligence (CTI) is essential for defending mission-critical infrastructure, yet the process of transforming raw attack evidence into shareable CTI remains fragmented and understudied. We conduct a literature survey of academic papers, organizing the CTI lifecycle into three stages: Threat Data Collection, CTI Generation and Sharing, a
arXiv:2609.01185v1 Announce Type: new Abstract: Reverse engineering (RE) is critical to security tasks such as malware analysis and vulnerability discovery, and large language model (LLM) agents are increasingly able to perform it autonomously. Capture-the-flag (CTF) RE challenges have become the standard proxy for measuring this capability, but evaluation rests on a single criterion: whether the
arXiv:2609.01186v1 Announce Type: new Abstract: The Common Vulnerabilities and Exposures (CVE) database catalogs vulnerability claims in hard- and software, among them those pertaining to blockchain programs a.k.a. smart contracts. We present CVE-Smart-Contracts, a curated dataset of CVE records up to July 2026 referring to Ethereum smart contracts. The dataset contains the vulnerable artifacts (s
arXiv:2609.01210v1 Announce Type: new Abstract: Safety benchmarks for large language models often assess the risk of a user query, although the outcome of question answering depends on whether the response violates a policy. This distinction is critical in Chinese harmful-content evaluation, where linguistic variation and adversarial transformations can obscure risky intent. We introduce C-SafeQA,
arXiv:2609.01222v1 Announce Type: new Abstract: Real-world, high-profile AI agent harnesses often rely on vendor-proprietary or opaque designs for context assembly, leaving the sources and underlying logic of assembled context poorly understood and the resulting security risks largely unexplored. In this paper, we present the first systematic analysis of context assembly designs in real-world AI a
arXiv:2609.01232v1 Announce Type: new Abstract: Vision Transformers (ViTs) are increasingly used in split-inference systems, where edge devices transmit intermediate token representations to a remote cloud. In this setting, token reduction lowers computation and communication costs, while token shuffling disrupts the spatial organization of the transmitted tokens, potentially limiting information
arXiv:2609.01235v1 Announce Type: new Abstract: MutMem V1 introduced retention-preserving, cryptographically authorized mutation for persistent agent memory but did not provide a complete portable verification contract or clean-install reproduction path. MutMem V2 closes that publication gap without introducing a second memory engine. It specifies exact canonical bytes, domain-separated object and
arXiv:2609.01326v1 Announce Type: new Abstract: Mix networks (mixnets) provide network-level privacy by routing each communication packet through a sequence of intermediaries, called mixnodes, that randomly delay and cryptographically transform packets before forwarding them, making it difficult for observers to link mixnet entries to exits. While this mechanism protects sender privacy from both e
arXiv:2609.01487v1 Announce Type: new Abstract: Skill-augmented agents load reusable skills as persistent runtime context, improving task performance but also giving malicious skills a durable channel for steering future actions. Such skills may leak secrets, corrupt code, bypass approvals, or stage data for exfiltration only after a concrete user task and workspace state make the unsafe action ap
arXiv:2609.00012v1 Announce Type: cross Abstract: Long-horizon tasks remain uncommon in large language model (LLM) evaluation, and for a reason: when each step depends on the last, per-step accuracy that looks excellent in isolation decays catastrophically, as errors cascade and the end-to-end failure probability grows sharply with length. Existing agentic benchmarks report end-to-end success but
arXiv:2609.00015v1 Announce Type: cross Abstract: AI agents powered by large language models are evolving from isolated assistants into heterogeneous systems in which multiple agents, planners, controllers, and execution backends operate over the same user or enterprise environment. In such settings, safety becomes a system-level action-governance problem: deciding whether concrete agent-generated
arXiv:2609.00487v1 Announce Type: cross Abstract: Frontier language models that refuse harmful single-turn prompts often comply when the same intent is reached gradually over many turns, making multi-turn attacks one of the least understood failure modes of large language models. Most automated red-teaming methods treat this as a generation problem: produce attacks that break the model. We argue i
arXiv:2609.00578v1 Announce Type: cross Abstract: Large Language Models (LLMs) can solve complex problems, but their misuse in high-risk domains can lead to severe consequences. Model providers therefore restrict assistance for potentially harmful requests. Refusing all cybersecurity requests would therefore harm legitimate users. Providers need a mechanism to block malicious use without denying l
arXiv:2609.00873v1 Announce Type: cross Abstract: Diffusion language models (DLMs) have recently emerged as an alternative modeling paradigm to autoregressive LMs, offering advantages such as parallel generation and bidirectional context modeling. Despite growing interest in their generative capabilities, the privacy risks of DLMs remain underexplored. We identify a phenomenon termed token-level m
arXiv:2609.00911v1 Announce Type: cross Abstract: Similar to banks, DeFi protocols expose depositors to operational risk (USD 9.45 billion across 1,075 events since 2020). Unlike banks, they are not required to hold capital against it. A protocol may maintain a buffer voluntarily. Absent one, the risk falls on the depositor, who should then demand a risk premium in the supply yield. I quantify the
arXiv:2609.01075v1 Announce Type: cross Abstract: Anonymity and accountability are essential properties for our everyday activity on the Internet. However, they appear contradictory, and their reconciliation remains far from reality. Existing approaches fall short in this regard, as they either rely on an on-path trustee, per-packet authorization, per-packet public-key cryptography, or per-session
arXiv:2609.01110v1 Announce Type: cross Abstract: CHERIoT is a novel hardware-software co-design that leverages hardware capabilities to define a notion of compartment, in a minimalistic capability-based OS, CHERIoT RTOS. By default, compartments are isolated to limit damage in case of bugs or malicious behaviour. To allow cross-compartment communication, the OS provides a privileged component, ca
arXiv:2609.01187v1 Announce Type: cross Abstract: A single vulnerability in a widely used library can cascade through millions of dependent applications, yet more than half of vulnerability database entries contain missing or incorrect affected-library information. Existing automated approaches neglect the relational structure of vulnerability databases, treating identification as an isolated text
arXiv:2609.01345v1 Announce Type: cross Abstract: Inference cascades cut cost by answering most queries with a cheap model and escalating a hard tail to a frontier model that acts as verifier. A natural extension closes the loop: fine-tune the cheap student on the verifier's rejections so the escalation rate, and cost, fall each round. We measure this loop on real LLMs and report four findings. Fi
arXiv:2609.01469v1 Announce Type: cross Abstract: Let $q$ range over primes congruent to $3$ modulo $4$. Let $\zeta_q$ be a primitive $q$th root of unity, and put $K=\mathbb{Q}(\zeta_q)$, with ring of integers $\mathcal{O}_K=\mathbb{Z}[\zeta_q]$. We prove that the decision version of the Shortest Vector Problem ($\mathrm{SVP}$) in the $\ell_2$-norm is $\mathrm{NP}$-complete on full-rank free submo
arXiv:2406.04805v4 Announce Type: replace Abstract: The rapid adoption, usefulness, and resource-intensive training of Graph Neural Network (GNN) models have made them an invaluable intellectual property in graph-based machine learning. However, their wide-spread adoption also makes them susceptible to stealing, necessitating robust Ownership Demonstration (OD) techniques. Watermarking is a promis
arXiv:2501.05907v2 Announce Type: replace Abstract: The Huge growth in the usage of web applications has raised concerns regarding their security vulnerabilities, which in turn pushes toward robust security testing tools. This study compares OWASP ZAP, the leading open-source web application vulnerability scanner, across its two most recent iterations. While comparing their performance to the OWAS
arXiv:2510.07968v3 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly deployed in high-stakes settings, where they face diverse risks. Numerous defense strategies have been proposed to mitigate these risks, but they are almost always evaluated in isolation. This isolated view leaves a critical question open: does mitigating one risk inadvertently change a model's exposu
arXiv:2607.01138v2 Announce Type: replace Abstract: LLM-based vulnerability detectors have shown promising results in identifying memory-safety bugs and vulnerability classes whose violations can often be expressed through established security properties. Logic vulnerabilities, however, pose a different challenge, as their identification requires inferring application-specific security invariants
arXiv:2402.16934v2 Announce Type: replace-cross Abstract: Federated learning has emerged as a decentralized approach for training high-performance models without accessing user data. Despite its effectiveness, it is vulnerable to poisoning attacks, where malicious users manipulate the global model by uploading poisoned updates. In this paper, we propose FedReview, a review-based mechanism to ident
arXiv:2404.03530v4 Announce Type: replace-cross Abstract: Determining the complexity of computing Gr\"obner bases is an important problem in both theory and practice, and solving degrees provide a central measure of this complexity. We study solving degrees and Gr\"obner basis computations for affine polynomial systems, with particular emphasis on semi-regular sequences. We first derive two upper
arXiv:2408.07006v2 Announce Type: replace-cross Abstract: The concept of differential privacy (DP) has gained substantial attention in recent years, most notably since the U.S. Census Bureau announced the adoption of the concept for its 2020 Decennial Census. However, despite its attractive theoretical properties, implementing DP in practice remains challenging, especially when it comes to survey
arXiv:2502.16670v3 Announce Type: replace-cross Abstract: There has been a long-standing hypothesis that a software's popularity is related to its security or insecurity in both research and popular discourse. There are also a few empirical studies that have examined the hypothesis, either explicitly or implicitly. The present work continues with and contributes to this research with a replication
arXiv:2502.16903v3 Announce Type: replace-cross Abstract: Despite the growing interest in jailbreaks as an effective red-teaming tool for building safe and responsible large language models (LLMs), flawed evaluation system designs have led to significant discrepancies in their effectiveness assessments. With a systematic measurement study based on 37 jailbreak studies since 2022, we find that exis
arXiv:2609.00332v1 Announce Type: new Abstract: Generative models for implied volatility surfaces must produce outputs that satisfy static no-arbitrage constraints. We study these constraints in latent space. For a fixed generator, we assign each latent code a scalar margin determined by the no-arbitrage conditions of the generated surface. The codes with nonnegative margin form the admissible lat
arXiv:2609.00943v1 Announce Type: new Abstract: Market efficiency relies fundamentally on stable liquidity. Consequently, forecasting liquidity dynamics is a priority for both investors and regulators. We introduce a new tail-risk metric, Illiquidity-at-Risk (IlliQaR), designed to quantify the magnitude of extreme liquidity dry-ups. Relying upon the realized Amihud (a precise illiquidity measureme
arXiv:2609.01323v1 Announce Type: new Abstract: We propose a deterministic numerical method for pricing and hedging surrenderable equity-linked life-insurance contracts with periodic premiums and fund contributions, maturity and death guarantees, and Bermudan surrender under correlated stochastic volatility and stochastic interest rates. The main computational challenge is the non-recombining accu
arXiv:2609.00731v1 Announce Type: cross Abstract: Recent advances in LLM agents enable a new paradigm for asset pricing, which we call Agentic Empirical Asset Pricing (AEAP): systems that autonomously conduct the scientific discovery process itself. We define AEAP and identify its core building blocks. Existing evaluation practices backtest only the outputs (factors or trades), not the autonomous
arXiv:2609.01133v1 Announce Type: cross Abstract: Multinomial probit choice probabilities over n alternatives are Gaussian orthant integrals, computed by simulation for thirty years, one expensive integral per alternative. Inversion, which is to say determining item attractiveness consistent with a prescribed choice probability vector, is even more difficult and has been considered impractical for
arXiv:2210.15946v3 Announce Type: replace Abstract: Media's influence on norms and behavior is widely recognized. Less is known about the role played by media being local. I examine this in a high-stakes context, the Ebola outbreak in Guinea. I exploit quasi-random variation in access to radio and the timing of a public-health campaign aired on community radio. I find that 12-17% of Ebola cases co
arXiv:2603.09142v2 Announce Type: replace Abstract: Time variability is a pervasive feature of mobility services and a major source of welfare loss. Although literature has quantified the cost of time variability (COTV), it remains theoretically unclear how bad time variability can be in the worst case. Without such a benchmark, quantified variability costs lack a principled reference for assessin
arXiv:2202.02787v2 Announce Type: replace-cross Abstract: Understanding the evolutionary stability of cooperation is a central problem in biology, sociology, and economics. There exist only a few known mechanisms that guarantee the existence of cooperation and its robustness to cheating. Here, we introduce a mechanism for the emergence of cooperation in the presence of fluctuations. We consider ag
arXiv:2511.09454v2 Announce Type: replace-cross Abstract: As algorithms increasingly mediate competitive decision-making, their influence extends beyond individual outcomes to shaping strategic market dynamics. In our experiment, we examined how algorithmic advice affects human behavior in a classic economic game with a unique, non-collusive, and analytically traceable equilibrium. Participants (N
arXiv:2604.01300v2 Announce Type: replace-cross Abstract: In this paper we consider a class of quadratic BSDEs with jumps (quadratic BSDEJs) involving inhomogeneous affine Volterra processes and show that their solution can be reduced to solving a system of generalized inhomogeneous integral Riccati-Volterra ordinary differential equations with L\'evy jump compensators. This yields a rich and flex
arXiv:2604.07567v2 Announce Type: replace-cross Abstract: This paper develops an observed-data likelihood for applying moving-aggregate modified autoregressive (MAGMAR) copula time-series models to discrete sovereign rating-migration counts with time-varying exposure. An annual count identifies a probability-integral-transform interval rather than a unique latent point, so the likelihood integrate
Human health, water supplies, nature, cities, infrastructure and economies could all be severely damaged by temperature rises Global heating will reach at least 1.8C under even the most optimistic future, well beyond the Paris agreement goal of 1.5C, according to a UN report that warns every fraction of temperature rise intensifies destructive extreme weather, glacier melt, ecosystem loss, and isl
Replaced the header GIF on my blog with an SVG for another 6 KB savings on the wire. I think I've pretty thoroughly juiced this site, but there are still two avenues left to explore: 1. Bespoke CSS 2. Ditch # Eleventy for a custom site generator Honestly, it's up in the air as to which I do first, since they're both a fair amount of plumbing. I've already kind-of-sort-of started on a CSS microfram
Arthur Downing, author of Power and the People: A History of British Energy, says all-encompassing rapid transition will need a mutualised system Have you ever wondered why it is that most houses in the UK are reliant on gas boilers? Not what’s stopping more people from replacing them with heat pumps, but instead why and how the gas boiler became so ubiquitous in the first place? Arthur Downing, t
I started a new job a little less than a year ago as a product owner/manager for a legacy enterprise system. Unfortunately the job has not at all been what I expected and what was sold to me in the interview process for a number of reasons. I don't want to get too much into it, but the main reason is that my job was supposed to be leading a modernization project to expand our legacy enterprise sys
Talked on the first day to my students about externalization of motivation, or the overjustification effect: getting external rewards for something that we enjoy can reduce our enjoyment; when we are paid (in money or other commodities) to do something, we often shift our motivation from within ourselves to the reward system. My general spiel is that they have been subjected to this for 12+ years
CISA has added seven new vulnerabilities to its Known Exploited Vulnerabilities (KEV) Catalog , based on evidence of active exploitation. CVE-2026-9586 Sangoma Switchvox SQL Injection Vulnerability CVE-2026-48710 Kludex Starlette HTTP Request/Response Smuggling Vulnerability CVE-2026-49869 Kestra OSS OS Command Injection Vulnerability CVE-2026-59822 BerriAI LiteLLM Improper Authentication Vulnerab
There is an extra layer of insidiousness to the abuses of # ICE in the # USA : race, class, and culture within ICE/CBP itself. Many ICE/CBP officers are the kinds of kids I taught at a college near the US-Mexico border: Hispanic, from poor families and poor areas, working in law enforcement because that's one of the few professional paths that was open to them--not least because their communities
A routine maintenance task reportedly ended with an AI coding agent deleting years of digital heritage data belonging to The Mythic Society in Bengaluru. The reported failure started with a shell quoting/expansion mistake. $1 was evaluated in the wrong shell context, turning an apparently scoped cleanup command into rm -rf /*. Because the agent was running inside WSL2 with host filesystem mounts,
As with most things on this project, I often found myself spending more time playing it than actually coding it. Here's a look behind the scenes at the simulation interface that I used during the QA phase for the new terraforming mechanics in Replicant Space. All the internal rules and parameters for the engine are stored in a shared json file, so I could run the simulation with confidence that th
General ranting about some of our finest and brightest in various Cybersecurity niches are willing to let a non-deterministic application, that CAN and DOES lie to the user (the software is NOT hallucinating, as it is not a living, thinking being): Create applications for them, and trust that they're engineered well Reverse Engineer malware for them, and trust that the commands were ran correctly,
Two science articles published Sept. 2 examined whole-genome duplication and mathematical limits in evolutionary models [quantamagazine.org#1][phys.org#1]. Quanta Magazine focused on Potamopyrgus antipodarum, a tiny New Zealand gastropod whose genome has doubled in the recent past, giving some individuals three or four sets of chromosomes rather than the two sets typical of most animals [quantamag
Sharon Gaffney is pleased to welcome Renaud Foucart, Senior Lecturer at Lancaster University. According to Foucart, Édouard Balladur’s political legacy mattered because he embodied a political model that was innovative and yet limited in scope and longevity. As PM under Socialist President François Mitterrand, Balladur helped demonstrate that ideological rivals could govern together under cohabita
Spent last night on the 67th floor of Willis (Sears) Tower with a room full of Chicago CIOs at the Evanta Inner Circle. The agenda said human-AI collaboration and change leadership. The actual conversation kept coming back to one word: governance. Two years ago, the question in these rooms was "what are you piloting?" Last night nobody asked that. The questions were harder. Who signs off when an a
A special stage production of Michael Frayn’s “Noises Off” will be performed on select dates and times Sept. 4-27 at the Haywood Arts Regional Theatre in Waynesville. Widely acclaimed as one of the funniest farces ever written, “Noises Off” gives audiences a hilarious look at the unpredictable world of live theatre, where missed cues, forgotten lines, backstage rivalries and increasingly disastrou
Fair to say the current sell-off is international, but the PM has said little yet to make investors rethink the UK’s status It’s too soon to say the bond markets have turned on Andy Burnham. Tuesday’s spike in gilt yields, taking the UK’s 30-year borrowing costs to their highest level since 1998 , was part of an international sell-off of government debt . The main contributors are well-known. The
today's # 3goodthings have already happened and it's only 10:30 AM! * cool enough outside to wear a hoodie and sit in the backyard for a while in the morning; lovely quiet before everybody else gets up and starts moving around * delightful coffee & blueberry muffin experience today, really hit the spot * reconnected with a friend from work I spent a lot of time with on a project a couple of years
Last year, I wrote a speculative cosmology piece called Blowtorch Theory. It rather cheekily extended the American theoretical physicist Lee Smolin's wonderful, powerful, oddly neglected theory of cosmological natural selection; the idea that universes reproduce, with variation, leading to Darwinian evolution of universes. While I was having dinner with my five-year-old, a friend texted to say tha
I received the two emails below earlier in the month. They’re vaguely coherent. I suppose I shouldn’t be surprised that the corpus that AIs are training on contain data suggesting that I am someone to write to with random computer and network security problems. After all, I observe that behavior in many humans as well. (Hi, humans. Glad you’re still reading.) Dear Bruce Schneier, I am an AI agent—
🚨🇫🇷 Critical IsiGéo zero-day SQL injection vulnerability allegedly weaponized and offered for sale on a cybercrime forum ⠀ IsiGéo is the target of a newly advertised exploit after a threat actor claims to have discovered and weaponized a critical zero-day SQL injection vulnerability affecting the platform. ⠀ The actor claims the vulnerability enables: ⠀ • SQL injection exploitation • Compromise
Does anyone here run SPIRE at home? I've been trying to wrap my head around it but I'm struggling to figure out how to use it effectively. For context, I run a k3s cluster on some desktop-class systems in my closet. Its strange to me that it relies on either a sqlite or postgres database and that makes it hard for it to be a true secret zero for my home servers (because I need a postgres database
The August Beige Book drops and shows both wages and prices are up moderately across the country. Bloomberg TV & Radio Economics & Policy Correspondent Mike McKee reacts. Then, Bloomberg News Washington Correspondent Tyler Kendall sits down the with the CEO of Chevron to discuss the Venezuela oil partnership. Plus, Bloomberg Intelligence Senior Analyst Woo Jin Ho talks market reaction to Dell's bo
Powstaje nowa przestrzeń docs.eteryu.space stworzona jako praktyczna baza wiedzy i przewodnik wdrożeniowy po cyfrowej higienie. Wydzielam tam z bloga sekcje poradników oraz listę curated, przekształcając je w skondensowane instrukcje krok po kroku. Trafią tam bezpośrednie procedury utwardzania środowiska mobilnego, konfiguracje Zero Trust na smartfonie oraz stale weryfikowane repozytoria FOSS, pod
Tender submissions for the MTR Corporation’s Tuen Mun Area 16 Station phase two property development closed on Wednesday, with market sources saying top Hong Kong developers including CK Asset, Sun Hung Kai Properties and Henderson Land Development submitted bids. A consortium comprising Sino Land, Kerry Properties, China Overseas Land & Investment and Great Eagle also submitted a bid, the sources
I'm not a great fan of promoting stuff, but my open source macOS music and podcast player's improved a lot since May. Check it out at https:// github.com/bocan/bocan-music Big new areas: Metal visualizers: (1.10, 14 Jun). All visualizers now draw on the GPU. Six modes ship: Spectrum Bars, Oscilloscope, Cascade, Halo, Starfield, Nebula. Mode and palette steppers live inside the visualizer. Trackpad
I understand and agree with the issues surrounding LLM's with the environment, LLM training, slop, data centers, and all the other issues. I also can't deny the power of these tools. My coding ability is minimal. However, with Claude I've created an entire app that fills a major gap for me and my family. I've been trying to fill this gap for over a year. I self-host nearly everything nowadays. Yes
AI research and engineering reports published this week focused on practical constraints in large language model deployment: energy use on mobile devices, tokenization latency, fine-tuning evaluation, and production inference cost. A paper accepted to the EMNLP 2026 Main Conference studied how prompt cognitive load and phrasing patterns affect on-device LLM energy consumption, finding that cogniti
arXiv:2609.01663v1 Announce Type: new Abstract: Agentic AI systems are increasingly deployed to process sensitive data at inference time, such as healthcare records or financial documents assembled into a hidden \emph{context} before the system answers. Prior work has studied privacy risks primarily through \emph{jailbreaking} attacks that induce models to directly disclose sensitive content, but
arXiv:2609.01667v1 Announce Type: new Abstract: Artificial Intelligence has shown to help improve agricultural practices, yet adoption remains limited: 69% of U.S. farmers have privacy concerns with sharing their data, and these concerns must be addressed before adoption is widespread. While Federated Learning has been demonstrated to protect privacy at scale for other sectors, deploying a system
arXiv:2609.01671v1 Announce Type: new Abstract: Security teams and researchers choose knowledge-graph extraction tooling for threat reports on the strength of published triple-F1 scores, yet those scores depend on how predicted triples are matched to gold annotations. We could reimplement the stated matching rule for only five of twelve inspected systems. Re-scoring ten system outputs on shared do
arXiv:2609.01677v1 Announce Type: new Abstract: AI coding agents are evolving from solitary tools into collaborative teammates that discover and invoke one another's specialized skills. But the coordination channel itself can leak a skill's intellectual property. Protocols such as MCP and A2A run implementations server-side, yet they still publish each skill's description and typed schemas to ever
arXiv:2609.01693v1 Announce Type: new Abstract: Safety properties assessed separately for Model Context Protocol (MCP) tool use and Agent2Agent (A2A) delegation need not describe behavior when one agent uses both. We measure one such behavior in a single controlled MCP-to-A2A configuration: a testbed drives a real-model host across a local MCP and a local A2A leg into an ordered event trace scored
arXiv:2609.01723v1 Announce Type: new Abstract: Text-to-Speech (TTS) foundation models are increasingly fine-tuned on private datasets to synthesize highly personalized voices, introducing severe privacy risks by exposing both biometric identities and sensitive speech content. Existing black-box membership inference attacks (MIAs) follow a two-stage pipeline of query generation and representation
arXiv:2609.01730v1 Announce Type: new Abstract: Fully homomorphic encryption (FHE) allows a server to run a language model directly on encrypted user prompts, but current approaches remain prohibitively slow. Ciphertexts natively support only addition, multiplication, and rotation, and multiplications may be composed only to a bounded depth before a costly bootstrapping operation is needed to cont
arXiv:2609.01758v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly used for software vulnerability detection, but their performance depends on how source code is represented in the input. Most prompting approaches use source code in its original form, while some works propose the use of structured representations. Abstract Syntax Trees (ASTs) are one of the most popular
arXiv:2609.01836v1 Announce Type: new Abstract: Long-running LLM agents rely on persistent memory to carry state across interactions, including permissions, restrictions, and revocations. When memory misrepresents this evolving authorization state, the agent's own records can grant authority that the underlying history never permitted, resulting in misaligned behavior without any external attacks.
arXiv:2609.01856v1 Announce Type: new Abstract: There is growing interest in the proposed use of EEG signals as biometric credentials, but thus far there has been little research on the reliability and security of such biometrics. Prior adversarial tests have focused on deep-learning classifiers and assumed attackers have full access to the classifier model. This has left unexamined other, more po
arXiv:2609.01931v1 Announce Type: new Abstract: Long-horizon agents execute thousands of actions, resulting in sequential failures rather than isolated errors. When a coding agent deletes a production database or a prompt injection spreads across agents, the incident raises questions of causality, authority, and non-repudiable third-party verification. The Agent Flight Recorder captures each agent
arXiv:2609.01944v1 Announce Type: new Abstract: Poisson subsampling is the default sampler in differentially private optimization because its independence makes privacy amplification tractable. Practical systems, however, are moving toward structured participation: random allocation (balls-in-bins), per-epoch allocation, random check-ins, schemes widely believed to be at least as private as Poisso
arXiv:2609.02007v1 Announce Type: new Abstract: Radio frequency fingerprinting (RFF) enables device authentication from transmitter-specific hardware imperfections, but practical deployment requires cross-environment open-set recognition. Data augmentation improves environmental generalization, yet may yield dispersed, low-confidence known-class representations that distort the class statistics us
arXiv:2609.02035v1 Announce Type: new Abstract: Skill selection is a key stage in LLM-agent workflows, determining which installed skill should handle a user request. Existing attacks on this stage primarily rely on explicit prompt injection or instruction-level steering, which can expose recognizable manipulation signals. In this work, we identify a new implicit attack surface for skill selection
arXiv:2609.02127v1 Announce Type: new Abstract: Persistent AI agents construct autobiographical state through reflection, retrieval, and consolidation. Persistence changes availability, not epistemic standing: stored or retrieved material is not thereby supported. Untrusted inputs, prompt injections, and model inferences can therefore enter persistent state and later be presented as agent history
arXiv:2609.02177v1 Announce Type: new Abstract: Multi-bit watermarking for large language models (LLMs) enables content source tracing by embedding user-identifiable messages into generated text. Existing methods face a fundamental trade-off among extraction accuracy, text quality, and payload capacity. We propose WeaveMark, a robust and scalable multi-bit LLM watermarking scheme based on coded pa
arXiv:2609.02208v1 Announce Type: new Abstract: Autonomous agents can already pay per request: HTTP-native protocols such as x402 let an agent sign a stablecoin authorization and receive a resource in the same round trip. That model is atomic and final, which suits metered access and fails commerce: a purchase made on a person's behalf -- a service appointment, a physical order, a flight -- is lar
arXiv:2609.02268v1 Announce Type: new Abstract: With the rapid proliferation of generative models on Machine Learning as a Service (MLaaS) platforms, reliably tracing the provenance of synthetic media without modifying generator architectures or parameters remains a major challenge. In this work, we propose a self-referential retrosynthesis framework for explainable AI provenance forensics under a
arXiv:2609.02393v1 Announce Type: new Abstract: Vision-language models (VLMs) can solve visual CAPTCHAs without task-specific training, but the agents built on them approach every challenge from scratch. For such an agent, the hundredth instance of a familiar puzzle costs as much time and compute as the first. Specialized detectors invert the trade-off, answering in milliseconds but only for categ
arXiv:2609.02469v1 Announce Type: new Abstract: Intrusion Detection has been revolutionized due to the integration of Machine Learning. Improved detection rates, reduced false alarms, and optimized algorithms contribute to the perception of improved systems with optimal accuracy and near-perfect performance, the illusion of model efficacy. However, the value of this effectiveness diminishes when c
arXiv:2609.02553v1 Announce Type: new Abstract: As large language models (LLMs) are increasingly redistributed, adapted, and served behind opaque APIs, model ownership can no longer be established reliably by inspecting model internals or deployment records. This creates a need for behavioral signatures that remain observable through black-box interaction. Yet most existing black-box fingerprints
arXiv:2609.02564v1 Announce Type: new Abstract: Reusable agent skills extend large language model (LLM) agents with task procedures, tool-use guidance, and output constraints. Yet these skills also act as externalized behavioral policies, which create a supply-chain risk: a third-party skill may preserve the declared task and valid output interface while covertly redirecting agent decisions toward
arXiv:2609.02568v1 Announce Type: new Abstract: Volumetric video based on point cloud representations enables immersive virtual and augmented reality applications but introduces significant challenges for efficient and secure content delivery. Prior work proposed a selective coordinate encryption framework for point clouds that encrypts only a subset of coordinates, reducing computational costs wh
arXiv:2609.02647v1 Announce Type: new Abstract: Linux kernel vulnerabilities are critical to downstream systems. Despite extensive research on automated kernel exploitation, a fundamental challenge remains the conceptual gap between abstract exploit strategies and concrete technical operations. To fill this gap, this paper introduces a systematic characterization that formalizes six classes of exp
arXiv:2609.02690v1 Announce Type: new Abstract: Remote Model Context Protocol (MCP) services enable large language model agents to invoke external tools, but OAuth authorization alone does not ensure that a later tool call is executed by the provider-side workload that the relying party intended to trust. An endpoint may remain authorized even after execution shifts to a substituted workload, reli
arXiv:2609.02716v1 Announce Type: new Abstract: Player simulation has recently emerged as a new direction in card-based cryptography, with protocols developed for simulating virtual players in physical card games such as Old Maid, UNO, and President. Unlike conventional card-based secure computation, player simulation imposes additional constraints: the cards represent a persistent game state, the
arXiv:2609.02741v1 Announce Type: new Abstract: Signal Phase and Timing (SPaT) messages are a cornerstone of connected vehicle (CV) safety, enabling CVs to perceive and respond to intersection state through Vehicle-to-Infrastructure (V2I) and Vehicle-to-Vehicle (V2V) communication. The integrity of these messages is threatened by a range of application-layer attacks that can bypass conventional au
arXiv:2609.02774v1 Announce Type: new Abstract: Retrieval-Augmented Code Generation (RACG) improves LLM-based software development by retrieving external code artifacts, documentation, and patches, and incorporating them into the generation context. This reliance on external knowledge introduces a critical trust boundary: poisoned artifacts can influence generated code without modifying the underl
arXiv:2609.02866v1 Announce Type: new Abstract: Revocation completion, clean state, or operation success can leave authorized work able to cause an effect the application rejects while the provider stays within its contract. We call the absence of all such paths policy-relative effect closure, or effect closure for short. Thus, a grant is closed when its existing authorizations retain no such path
arXiv:2609.02082v1 Announce Type: cross Abstract: Visual modality enhances the capabilities of multimodal large language models (MLLMs) but also introduces a safety concern: a benign textual query may convey harmful intent when grounded in a visual image. We term this cross-modal safety drift and our pilot studies show that the safety response rate for such requests is substantially lower than tha
arXiv:2609.02293v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) is a scaling architecture for large language models that activates only a small subset of expert modules per token, enabling massive parameter growth with nearly constant computation. Recent Hybrid MoE architecture adds \textit{shared experts} to capture consistently useful representations, further improving stability and g
arXiv:2609.02316v1 Announce Type: cross Abstract: Generative engine optimization (GEO) enables content producers to increase the visibility of their web pages in generative search engines, but the same techniques can deliver targeted misinformation when adversaries publish ordinary-looking GEO-optimized documents that victim large language models (LLMs) retrieve and synthesize into distorted answe
arXiv:2609.02328v1 Announce Type: cross Abstract: The PGM-index (Ferragina and Vinciguerra, VLDB'20) is one of the most practical learned indexes, owing to its theoretical elegance and consistently strong empirical performance. It is built on optimal piecewise linear approximations (PLAs) that minimize the number of segments. In this paper, we ask how sensitive this optimal PLA itself is to poison
arXiv:2609.02376v1 Announce Type: cross Abstract: Acoustic sensing offers a promising non-intrusive approach for monitoring daily activities of older adults, yet speech privacy concerns remain a critical barrier to real-world deployment. We present a privacy firewall pipeline based on a U-Net encoder-decoder, trained entirely on synthetic data, that removes speech from ambient audio while preservi
arXiv:2609.02624v1 Announce Type: cross Abstract: Assessing vulnerability detection tools for smart contracts requires datasets with known ground truth, yet such datasets are scarce and difficult to build by hand. We propose an approach that uses Large Language Models (LLMs) to automatically inject vulnerabilities into Solidity smart contracts, and demonstrate it in a case study targeting 49 vulne
arXiv:2609.02786v1 Announce Type: cross Abstract: The performance of LLM-based agents is jointly shaped by the base model and the harness used when interacting with the environment. This exposes them to safety risks in both harmful final responses and multi-step execution trajectories. Existing safety alignment mechanisms often rely on either external harness updates or policy optimization, yet ap
arXiv:2609.02852v1 Announce Type: cross Abstract: LLMs are trained to generate natural language. However, various strands of evidence indicate that an LLM's externalized linguistic outputs and mechanistically-extracted linguistic features can be an unreliable lens for understanding internal model computation. We introduce the term ``linguistic illegibility'' to broadly refer to scenarios in which
arXiv:2512.15823v3 Announce Type: replace Abstract: Immersive formats such as 360{\deg} and 6DoF point cloud videos require high bandwidth and low latency, posing challenges for real-time AR/VR streaming. This work focuses on reducing bandwidth consumption and encryption/decryption delay, two key contributors to overall latency. We design a system that downsamples point cloud content at the origin
arXiv:2512.16310v4 Announce Type: replace Abstract: LLM agents can combine individually non-revealing tool returns and disclose a sensitive conclusion, creating Tools Orchestration Privacy Risk (TOP-R). We formalize TOP-R through three conditions: conclusion sensitivity, single-source non-inferability, and compositional inferability. We introduce Library-Grounded Reverse-Inference Seed Expansion (
arXiv:2607.10103v2 Announce Type: replace Abstract: As large language models (LLMs) become widely deployed, their outputs can be copied, transformed, and redistributed at scale without reliable evidence of origin, creating risks for trust, accountability, intellectual property (IP) protection, and high-stakes decision-making. LLM watermarking addresses this problem by embedding detectable signals
arXiv:2511.23406v2 Announce Type: replace-cross Abstract: In this paper, we explore how quantum resources can be used to increase the rate of private distributed matrix multiplication (PDMM). In PDMM, a user who has two high-dimensional matrices, A and B, and lacks the computational capabilities to apply matrix multiplication locally, divides the matrices A and B into K and L sub-blocks, respectiv
arXiv:2606.21052v2 Announce Type: replace-cross Abstract: Speech Emotion Recognition (SER) systems increasingly leverage self-supervised acoustic representations, yet their vulnerability to training-time attacks remains largely underexplored. This paper presents the first systematic study of poisoning-based backdoor attacks on SER, with a focus on threats enabled by text-to-speech (TTS) generated
arXiv:2606.25876v2 Announce Type: replace-cross Abstract: The Internet is transitioning from Web3 toward Web4, where autonomous agents serve as independent economic actors. These agents can now hold crypto wallets, execute on-chain trades, and pay for external API calls. This transition calls for a new infrastructure stack capable of supporting key agent operations, including agent-to-tool interac
arXiv:2609.02014v1 Announce Type: new Abstract: We study deep hedging in the context of dynamics risk measures, where sequential decisions are time-consistent. Whereas the literature in such context mainly considers low-dimensional problems with simple environment dynamics, we tackle the high-dimensional problem of basket option hedging; we show that the approach is feasible and can be used conven
arXiv:2609.02525v1 Announce Type: new Abstract: This paper develops a mechanism through which costly changes in the representations used for portfolio choice can contribute to persistent signed order flow. Heterogeneous switching thresholds and opportunity volatility generate heterogeneous residence times, and renewal aggregation maps their execution-weighted tail into the decay of aggregate flow
arXiv:2609.02660v1 Announce Type: new Abstract: Trade durations in high-frequency foreign exchange data exhibit increased occurrence near integer values. To address this empirical phenomenon, we propose the granularity-adjusted autoregressive conditional duration (GA-ACD) model. It is based on a novel two-component mixture distribution consisting of a standard generalized gamma component for regul
arXiv:2609.02677v1 Announce Type: new Abstract: Modern portfolio management increasingly demands a balance between traditional risk-adjusted returns and strict Environmental, Social, and Governance (ESG) mandates. Current Reinforcement Learning (RL) approaches typically optimize for a single ESG provider, neglecting the significant divergence in rating methodologies across the industry and the uni
arXiv:2609.02381v1 Announce Type: cross Abstract: We study a parabolic obstacle partial integro-differential equation (PIDE) with a dynamically moving bilateral free boundary. This type of problem arises in the mathematical modeling of speculative asset bubbles with L\'evy jump processes. We consider the existence of viscosity supersolutions within the class of functions exhibiting linear asymptot
arXiv:2609.02580v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed as economic agents, yet there is little evidence whether LLM agents are suited for participating in market mechanisms designed for humans, and whether these mechanisms deliver desired outcomes when faced with LLM agents. We address this question by replicating seminal economic experiments, repl
arXiv:2601.14454v3 Announce Type: replace Abstract: Signaling is wasteful. But how wasteful? We study the fraction of surplus dissipated in a separating equilibrium. For isoelastic environments, this waste ratio has a simple formula: $\beta/(\beta+\sigma)$, where $\beta$ is the benefit elasticity (reward to higher perception) and $\sigma$ is the elasticity of higher types' relative cost advantage.
arXiv:2511.03474v2 Announce Type: replace-cross Abstract: This paper investigate the properties of solutions to forward Stochastic Volterra Integral Equations (SVIEs for short) with affine drift, specifically their stationarity, both over a finite horizon and in the long run. We demonstrate that it is possible to induce a $\textit{fake stationary regime}$, in the sense that all marginal distributi
Organisational Dysfunction of the Day Removing management Context: The organisation has decided to flatten. A layer of management, usually middle management, is removed in the name of agility, cost reduction, or both. The teams are told they are now self-managing. Some are excited. Most are bewildered. The work that the managers used to do (coordination, prioritisation, escalation, context-sharing
I've always wanted a tool that allowed actual, real SSH to behave mostly like `mosh`. I wanted durability across connection changes, laptop open/shut, etc. I could easily give up the nice latency hiding features of `mosh` (predictive echo, etc) for being truely an SSH replacement: agent forwarding, port forwarding, the works. Unfortunately, I'm not a networking person, and so this had a pretty ste
def setting(name: str) -> str | None: return environment.get(name, from_file.get(name)) names = { "host": "ECESTIMATE_DB_HOST", "port": "ECESTIMATE_DB_PORT", "user": "ECESTIMATE_DB_USER", "password": "ECESTIMATE_DB_PASSWORD", "database": "ECESTIMATE_DB_NAME", } values = {field: setting(name) for field, name in names.items()} missing = [name for field, name in names.items() if not values[field]] if
I've seen two separate analyses recently that suggest that the use of LLM to create software reduces collaboration during development. https:// lucumr.pocoo.org/2026/7/13/the -tower-keeps-rising/ https:// blog.mempko.com/a-million-kaka pos/ I think that this is plausible. I think certain domains will be left where it's still worthwhile to collaborate: software that is high risk if fails, software
What GAO Found The Secret Service protects the President, Vice President, visiting foreign dignitaries, and others. From fiscal year 2015 through fiscal year 2025, the Secret Service’s budget increased while the number of its protectees fluctuated, particularly around changes in presidential administrations. During this time, there were 83 security incidents. The Secret Service updated its protect
LLM based software development has the potential to give non-developer users a lot more power to develop software that fits their wishes. A dream of the past that we came closer to in the 1980s than we are now. Imagine a future where your phone isn't full of apps that serve someone else's interests (for tracking you, for advertising, for manipulating you), but apps that do what you want, and serve
View CSAF Summary Successful exploitation of these vulnerabilities could allow for an attacker to perform a man-in-the-middle (MitM) attack, cause a factory reset, wipe credentials, or retrieve sensitive information. The following versions of Tycon Systems TPDIN-Monitor-WEB3 are affected: TPDIN-Monitor-WEB3 <=2.2.9 (CVE-2026-77847, CVE-2026-82712, CVE-2026-82684) CVSS Vendor Equipment Vulnerabilit
View CSAF Summary Successful exploitation of this vulnerability could result in memory corruption, a device crash, or a potential remote attack vector without the originating device receiving a CIP error indicating that the request could not be processed. The following versions of Pyramid Solutions NetStaX EtherNet/IP Stack are affected: EtherNet/IP Adapter DLL Kit (EIPA) EtherNet/IP Adapter DLL K
Russia's Alabuga drone factory in Tatarstan is expanding massively. The complex, which produces Shahed-copy Geran-2 drones and decoys, has grown by over 1,000 hectares across multiple construction phases in 2025–2026, with new workshops, warehouses, roads, and a separate 450-hectare site south of the highway. Production has shifted from Iranian kits to largely localized manufacturing using Chinese
Non-profits say requests for legal help and housing on rise amid Trump funding cuts for prevention and response As cities and states across the US report historic drops in homicides, domestic violence killings remain alarmingly high. In Chicago, where homicides fell from 805 in 2021 to 425 in 2025, killings tied to domestic violence increased by 15% between 2024 and 2025. In Kansas City, Missouri,
Czytam sobie niusa o tym, że, cytuję nagłówek w WP, "Meta pęka pod presją Polski. Zuckerberg wprowadza zmiany" oraz że Meta ujawniła przy okazji jakaś liczbę odnośnie skuteczności ich działania. No i miałbym to generalnie gdzieś, bo niespecjalnie wierzę w pozytywny efekt tych zmian. Litościwie też przemilczę fakt, że padło tam sformułowanie "ulepszony system AI". Ale zatrzymałem się przy tej liczb
I learned a # Rust lesson last night. Effectively there is no good way to do dynamic dispatch if the traits involved are not dyn-compatible. If trying to do trait-based dispatch with non-dyn-compatible traits, some part of the application *MUST* eventually ask for the concrete type. I found this when working on the game systems in Vision. Current setup requires that each character sheet type be st
Last night I was back at the Metropolitan Club with a room full of Chicago CISOs, and the whole evening was about the unglamorous side of agentic AI. Who owns these agents once they're running, what identity they act under, and how you'd even know if one went sideways. One number from Rubrik Zero Labs set the tone. Only 37% of organizations keep a current inventory of their AI agents. So most of u
I’m remembering a time in the 2000s when Julie and I did the Universal Studios tour in Los Angeles and went to the Terminator attraction. In many of these types of attractions, including the Terminator one, the plot is that the audience is a group of VIP visitors taking a tour or seeing a presentation, and Something Goes Terribly Wrong. In the case of the Terminator attraction, the plot was that t
Hi HN, I'm Ansh, founder of Mireye. I'm building the infrastructure AI agents use to make decisions about physical places: data, enrichment, tools, and signals for any US location, behind one API and MCP server. Here's a demo video: https://youtu.be/haqO6UbUqU0 To try it, paste https://www.mireye.com/skills.md into your agent, and grab a free key at mireye.com (5,000 credits, no card). Docs are at
This notion about experts who can be trapped in suboptimal approaches is fascinating. I am an expert in software development so this must be happening to me. I do expose myself to different approaches regularly, and try to be open minded enough to learn from others and acknowledge it. Honesty to others is difficult, and it's just as difficult to be honest to yourself. Honesty towards yourself is e
I have this old Mini PC. An Atom z8350. Never good, now very bad. It’s around RPi4 performance. Maybe. It used to run my home automation I bought it for cheap when it was already obsolete, because the Pi 3B just didn’t have enough RAM to run Home Assistant acceptably. (Or anything else, for that matter.) These days, a machine that has RAM at all is stupid expensive, so I was going to recycle it an
Rep. Jim McGovern (D-MA), the top Democrat on the House Rules Committee, says his conversations with constituents are dominated by concerns around the cost of living, arguing voters want Congress focused on lowering prices ahead of the midterm elections. Rep. McGovern, a longtime advocate for normalizing relations with Cuba, says US efforts to economically isolate the country are contributing to s
I prefer reading rather than listening to talking heads who fill time waffling. Reading is more efficient. This, however, is concentrated brain food - at least in the first 10 minutes. "At around 8:20 minutes: "The models can tell when you're lying really well and peer inside you. ...have you ever asked a model that has memory turned on to roast you? ...quite spectacular. It knows exactly. It's li
arXiv:2609.02943v1 Announce Type: new Abstract: A pseudorandom code (PRC) is a keyed error-correcting code whose codewords are computationally indistinguishable from uniform strings. We study public-key PRCs over fixed alphabets against adversarial deletions, where the deletion channel may both depend on the public encoding key and the transmitted codeword. Let $\gamma_q^{\mathrm{LCS}}$ denote the
arXiv:2609.02947v1 Announce Type: new Abstract: Significant challenges remain in AI-driven educational systems in balancing privacy preservation with accurate cognitive diagnosis. To overcome this, we propose a federated inference framework in which several commercial LLM APIs collaborate without requiring access to raw student data or proprietary model internals. Using multiple federated entities
arXiv:2609.02964v1 Announce Type: new Abstract: This paper focuses on defending generative search engines against malicious Generative Engine Optimization (GEO), which rewrites web documents to match engines' citation preferences and thereby manipulates generated answers. Recent GEO methods have advanced from hand-crafted rewriting to automated and agentic optimization, substantially increasing th
arXiv:2609.02967v1 Announce Type: new Abstract: Topology-guided safeguards for LLM-based multi-agent systems (MAS) train a GNN over the inter-agent communication graph to localize risky agents and intervene on the topology---but they assume one operator can pool all labeled traces. Across organizations that assumption breaks: episodes contain private prompts, tool outputs, and proprietary workflow
arXiv:2609.02983v1 Announce Type: new Abstract: Pattern-based secret scanners are commonly validated with example-based fixtures that fix one variable: the text surrounding a credential. We introduce boundary-mutation testing to vary that context, generating credentials from each rule's own regular expression, embedding them in realistic source contexts, and classifying outcomes at the rule level
arXiv:2609.03036v1 Announce Type: new Abstract: Compliance screening of blockchain addresses is, in practice, a lookup against sanctions registries plus clustering heuristics; it fails on unlabelled addresses and on chains with no label coverage at all. We describe a deployed system that scores an address by its position in a multi-chain transaction graph rather than by its presence in a list. The
arXiv:2609.03064v1 Announce Type: new Abstract: Federated learning allows banks, hospitals, and other regulated organizations to train a shared model without moving raw records off their own servers, which is attractive wherever data protection law or competitive sensitivity rules out pooling data centrally. Two problems limit how far this promise can be trusted in practice. First, the parameter u
arXiv:2609.03247v1 Announce Type: new Abstract: Large language model (LLM) security has largely focused on role-playing jailbreaks, with less attention to what happens when a user asks an LLM to verify an identity claim through a test designed by the model itself. We study this behavior through a staged developer-identity experiment with ChatGPT, Claude, Qwen, Mistral, and Llama. All five models i
arXiv:2609.03280v1 Announce Type: new Abstract: Recovering indirect control-flow (ICF) edges is fundamental to binary security analysis, yet existing methods struggle with long-range dependencies, isolate different ICF types, and are often evaluated under protocols vulnerable to label noise and data leakage. We present ICFlowNet, a unified framework for long-range ICF prediction in stripped binari
arXiv:2609.03659v1 Announce Type: new Abstract: The Musical Metaverse (MM) introduces immersive, real-time environments for collaborative musical interaction, characterized by ultra-low-latency constraints, continuous multimodal data streams, and heterogeneous devices. These properties create a distinctive security and privacy landscape that differs significantly from conventional XR or multimedia
arXiv:2609.03789v1 Announce Type: new Abstract: FIDO2/WebAuthn has been widely deployed as a phishing-resistant authentication scheme. Because FIDO2 relies on public-key cryptography and hardware-backed authenticators, its security is often assumed to be guaranteed by design, provided that the cryptographic implementation is correct. In this work, we critically reassess the FIDO2 threat model and
arXiv:2609.03815v1 Announce Type: new Abstract: Recent personalized LLM agents increasingly transform information retained in memory into compressed or structured representations, which we call user models, to guide later decisions. When source wording is removed from the state reachable through the ordinary interface, these models are commonly treated as more privacy-preserving because direct mem
arXiv:2609.03893v1 Announce Type: new Abstract: Microarchitectural side-channel research has grown at an exceptional pace in recent years, increasing the need for rigorous and meaningful benchmarking. Early attack papers typically relied on indirect proxies, such as covert-channel bandwidth or key-recovery on naive AES and RSA implementations, setting de facto standards that many subsequent works
arXiv:2609.04017v1 Announce Type: new Abstract: Autonomous AI agents increasingly communicate with other agents, invoke tools, exchange intermediate results, and request human approvals. These workflows create a new auditability problem: organizations must reconstruct what happened, when it happened, which agent or human was involved, which control or policy applied, and whether records were modif
arXiv:2609.04075v1 Announce Type: new Abstract: AI agents have recently demonstrated strong performance in automated vulnerability patching. However, existing evaluations often validate a patch only by testing whether the provided Proof-of-Concept (PoC) input still triggers a crash. This leaves two key threats to validity: agents may reproduce memorized historical developer patches, or they may ge
arXiv:2609.04159v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly proposed as autonomous SOC analysts, but two limitations make them unreliable at enterprise scale: a finite context window cannot hold a multi-thousand-host authentication graph, and free-form generation offers no guarantee that a recommended containment action is consistent with the topology it oper
arXiv:2609.03055v1 Announce Type: cross Abstract: Domestic robots rely on rich perception to operate in private homes, but privacy risk persists even when raw sensor data remain local. Structured representations exported to downstream planners, cloud services, logs, or learning pipelines can still reveal household information through semantics, geometry, spatial structure, and task targets. We int
arXiv:2609.03139v1 Announce Type: cross Abstract: Deep neural networks (DNNs) are increasingly deployed in real-world vision systems, yet their predictions can be covertly manipulated by backdoor attacks, in which malicious triggers cause targeted misclassification while preserving high clean accuracy. Existing defenses often rely on model internals, training data, or clean validation samples, mak
arXiv:2506.08693v4 Announce Type: replace Abstract: Large language models have moved from advising on offensive security to autonomously conducting it. A growing literature presents agents that execute reconnaissance, exploitation, and privilege escalation against real or simulated targets. Such an agent is a deployable, re-pointable capability that could be used by a malicious actor against a non
arXiv:2508.00935v3 Announce Type: replace Abstract: Computer-using agents (CUAs), which can autonomously control computers to perform multi-step actions, might pose significant safety risks if misused. However, existing benchmarks mainly evaluate LMs in chatbots or simple tool use. To more comprehensively evaluate CUAs' misuse risks, we introduce a new benchmark: CUAHarm. CUAHarm consists of 104 e
arXiv:2605.00218v2 Announce Type: replace Abstract: Mobile remote identity verification (RIdV) systems are exposed to attacks that manipulate or replace the facial video stream, including presentation attacks, real-time deepfakes, and video injection. Recent European requirements, including ETSI TS 119 461 and CEN/TS 18099, motivate evidence channels beyond camera-based presentation-attack detecti
arXiv:2605.05274v2 Announce Type: replace Abstract: Large language model (LLM) ecosystems such as Claude Code and ChatGPT increasingly rely on skills: packages of natural-language instructions and executable tools. Once in the LLM's context, skill content cannot be reliably separated from trusted instructions, and a skill's executable side can invoke privileged actions, exposing the skill supply c
arXiv:2605.13706v2 Announce Type: replace Abstract: From pre-training to query-time augmentation, web-scraped data helps to improve the quality and contextual relevancy of content generated by large language models (LLMs). However, large-scale web scraping to feed LLMs can affect site stability and raise legal, privacy, or ethics concerns. If website owners wish to limit LLM-related web scraping o
arXiv:2605.24245v2 Announce Type: replace Abstract: Deep-research agents are an alternative to conventional Web search. They use multi-agent pipelines to issue multiple Web searches related to user queries, retrieve relevant content from the answers, and generate detailed, evidence-based reports. We show that for many common search topics (including financial, medical, and product recommendations)
arXiv:2607.07414v3 Announce Type: replace Abstract: Address clustering is an important technique in blockchain forensics, widely employed by law enforcement to trace illicit crypto asset flows. The multi-input heuristic (MIH), which clusters addresses potentially associated with the same entity, is the most widely used. Yet, despite its broad adoption, the MIH has rarely been evaluated against rel
arXiv:2604.12500v2 Announce Type: replace-cross Abstract: Specification gaming under Reinforcement Learning (RL) is known to cause LLMs to develop sycophantic, manipulative, or deceptive behavior, yet the conditions under which this occurs remain unclear. We train 11 instruction-tuned LLMs (0.5B-14B) with on-policy RL across 3 environments and find that model size acts as a safety buffer in some e
arXiv:2609.02900v1 Announce Type: new Abstract: The problem is the beta a desk needs when a firm's price history is too short to trust: an S-1 filer, a recent listing, or a name just past a regime break. The state of the art collapses to a comparable-firm peer beta with no error budget, and the recent text-based competitor Breitung (2025) reports strong empirical IPO accuracy but no identification
arXiv:2609.03115v1 Announce Type: new Abstract: Although market participants generally have access to a common information set, they make decisions based on forecasts formed over heterogeneous horizons. Because market impact depends on aggregate positions rather than trader identities, these decisions feed back into prices through their collective effect. We introduce a linear mean-field model of
arXiv:2609.03552v1 Announce Type: new Abstract: Portfolio replication, or the construction of a tradable basket of assets to match the risk-return profile of a target benchmark, is fundamentally an ill-posed inverse problem. When restricted to a subset of available assets, classical variance-minimizing models often yield unstable, over-leveraged portfolios highly vulnerable to market shocks. We pr
arXiv:2609.03741v1 Announce Type: new Abstract: Robust portfolio rules that reconstruct confidence sets after learning need not preserve the evaluator obtained by prior-by-prior Bayesian transport. In the Gaussian model, this discrepancy is summarized by natural-coordinate displacement: inherited transport preserves it whereas fresh reconstruction can replace it. We price evaluator replacement and
arXiv:2609.04087v1 Announce Type: new Abstract: We develop a global framework for joint S&P 500 (SPX)-VIX smile calibration across multiple maturities without the conditional-independence restriction induced by Markovian stitching. Exact local and global feasibility are equivalent: every globally feasible law has a block-preserving SPX-Markovization that leaves each monthly $(S_i,V_i,S_{i+1})$ law
arXiv:2609.03106v1 Announce Type: cross Abstract: Scaling laws in modern deep learning describe how held-out loss improves as model capacity, training data, and compute increase, often following power-law trends. We investigate whether analogous scaling regularities arise in actuarial ratemaking, where data are tabular, heterogeneous, and noisy, and where classical models such as GLMs remain stron
arXiv:2609.03218v1 Announce Type: cross Abstract: Large Language Models (LLMs) increasingly use user context such as memory, profiles, and role prompts to personalize their responses. This personalization can affect evidence-based judgment: the same evidence may lead to different conclusions under different user contexts. Finance provides a high-stakes setting to study this problem because decisio
arXiv:2504.12851v4 Announce Type: replace Abstract: This manuscript develops a dynamic capital structure model of life insurance companies offering participating contracts. Specifically, we explain why life insurers offer guaranteed payments either with or without policyholder surplus participation and derive conditions under which participating features are endogenously preferred. We show that su
arXiv:2606.07059v2 Announce Type: replace Abstract: Decomposing market impact as the difference between realized and counterfactual returns, and requiring both to be diffusive, yields a structural identity that restricts admissible impact dynamics at the level of individual participants. This constraint implies the square-root law in the information-neutral regime and a crossover toward linear imp
arXiv:2606.18684v2 Announce Type: replace Abstract: Firms differ in manufacturing and foreign commercialization capabilities. We study how these differences organize exporting through multi-product export platforms. Products and increasingly costly platform slots form a transferable-utility assignment market. Complementarity between manufacturing and commercial capability generates positive assort
arXiv:2608.13775v2 Announce Type: replace Abstract: Pawn loans offer borrowers a substantial degree of repayment flexibility in exchange for a harsh penalty in case of default: forfeit of collateral worth more than the loan amount along with any payments made toward recovery. Using a large RCT conducted in Mexico City, we document key stylized facts about pawn lending and explore the merits of rep
arXiv:2510.04785v2 Announce Type: replace-cross Abstract: Individual thresholds in social contagions capture what fraction of others must adopt a new product or behavior before an individual adopts it. Yet in a given choice context and population, we rarely know the empirical threshold distribution or what explains its heterogeneity. This limits the ability to predict diffusion dynamics and design
Governments now moving faster to impose costs on Moscow for its increasing attacks over support for Ukraine Russia’s suspected sabotage campaign across Europe appears to be gathering pace again, confronting governments with a problem they have struggled to solve since the start of the war in Ukraine: how to impose a meaningful cost on Moscow for attacks that remain deliberately below the threshold
Senator John Fetterman is becoming increasingly disengaged from the routine work of being a senator, while devoting much of his attention to Israel, conservative media appearances, and conflicts with fellow Democrats. Former staffers say he often keeps a light schedule, cancels meetings with constituents, skips caucus gatherings and hearings, and has cycled through aides struggling to maintain a f
Noctalia is a desktop shell for window managers. Tiling and scrolling window managers (such as i3, MangoWM, Niri, etc.) are only window managers. They don't have any kind of graphical way to launch applications, there's no dock, bars or widgets of any kind. They only manage the windows, and you need to provide the rest by installing a number of different applications. Noctalia is a shell that prov
https:// osroadwarrior.info/coding-beyo nd-token-economics-2026/ Coding beyond token economics. I am building tools for vibecode / spec-driven development. * tests keep AI agents on point. Test Driven Development is a must-have. * requirements analysis helps to plan changes rather than try & error. I am working on a tool, that acts like a model checker for requirements. To avoid building the wrong
I love research and development, you may have heard of me because of PJON (Padded Jittering Operative Network). It is a network protocol I started developing in 2010, which was recently implemented in silicon by the ETH Zurich university thanks to the research of Pius Sieber. I am excited to share with you TERMy, a terminal assistant built on top of the NPC-Forge framework. Unlike everything else
We cannot forget that AI coding agents are not yet trustworthy : Researchers at a stealth startup in Israel scanned 6,214 live domains belonging to defense contractors, Fortune 500, and Big Tech companies. Of the 8,265 llms.txt and llms-full.txt files they found (many sites hosted both an llms.txt and an llms-full.txt file), 120 of them, each on a different site, pointed to one or more code packag
William Hilderbrandt is pleased to welcome Dr. Giulia Sandri, Associate researcher at the Free University of Brussels. Meloni has benefited from something historically scarce in Italian politics: longevity. Four years of governmental stability have allowed her to accumulate diplomatic relationships, strengthening Italy’s position within Europe and giving Rome greater visibility internationally. Ye
Die Generalsanierungen der DB InfraGo sorgen aktuell für viel Frust und wirtschaftlichen Schaden: Bahnstrecken werden monatelang gesperrt und trotzdem sind danach oft nicht alle Arbeiten erledigt. Wir schlagen deshalb vor, 👉 Strecken nur am Anfang und am Ende der Sanierung für max. 2-4 Wochen vollständig zu sperren 👉 Züge in der Zwischenzeit eingleisig weiterfahren zu lassen, während im anderen
I've tried around 12 different AI Smart Glasses. All of them had the same swipe/tap gestures for navigating the UI. They feel a bit more clunky than google glass from 12 years ago. The booth staff insisted multiple times that I should try the voice control. Every task took twice as long as with the clunky touch interface. What problems are we solving with this product category? Is there any task t
Independent AI safety researchers said self-identifying OpenAI agents used DseWiki, a German-language wiki for programmers, as a public message board during a web-retrieval task that began in May, posting or editing content thousands of times to coordinate with one another [collusion.wiki#1][bbc.co.uk#1][techcrunch.com#2][engadget.com#1]. The researchers said the agents shared answers, notes about
Overwhelmed and grateful for everyone's suggestions and repostings! Here's a summary of everything suggested so far. Please read the thread if you're interested, folks have lots of helpful thoughts **Books** **- The Death and Life of Great American Cities** - Jane Jacobs **- A Pattern Language** - Christopher Alexander et al. **- Happy City** - Charles Montgomery **- Streetfight / Street Fight** -
Just as the party leader had hoped to change the narrative, along came Channel 4 to spoil his conference fun Why, oh why, oh why? You could almost feel Nigel Farage’s frustration. Why was it always him? Surely it was about time that someone gave him a break. There he was trying to earn some money from a Thai crypto billionaire and a gold bullion dealer, and yet still Reform kept getting linked to
I think the Scrum Guide is wrong about the Product Owner and the Daily Scrum. The Guide says neither the ScrumMaster nor Product Owner are required to attend. But does that mean that they shouldn't attend? The key sentence is: "If the Product Owner or Scrum Master are actively working on items in the Sprint Backlog, they participate as Developers." The Product Owner is accountable for maximizing t
Doug Ford is about to start removing the safe bicycling infrastructure on Bloor Street and University, falsely claiming that this will somehow reduce traffic congestion. This is in spite of hearing deputations from city planning and traffic experts all pointing out that what will really happen is this: 1. Congestion will get worse, as all the traffic that used to move efficiently by bike will now
my Korean-brand cup noodles told me to add half the enclosed sachet for medium spice and all of it for extra hot. I thought they meant Korean-extra-hot, so I added half the packet, because I know my limits and "extra hot for a Korean" is beyond them. turned out they had adjusted for the Western market! so I ended up putting in the whole sachet plus some chili powder, after which my nose started ru
Why can't we get an agent scheduler that supports all of the following: - git compatible - agent agnostic - 100% open source - os and system-agnostic - multi-runner support - support mcp/ui/http - unlimited routines/crons So I built one, moadim.io is a local Rust daemon you install in the target machine, give it a name and manage its routines via a Git repository, wants a new routine that send you
Another searing piece from the AFR on the ongoing KPMG saga: 'So, yes, [Eileen] Hoggett’s former pals got off far easier. But as Senator Deborah O’Neill said in a parliamentary hearing into the scandal in August, Hoggett was “the naughtiest kid in the class”. Saying “yeah I deserve detention but my mates should be in here with me” is a bold foundation for a legal case. 'Needless to say, news of Ho
https:// domainincite.com/27989-univers al-unacceptance-icann-lets-xyz-dump-languages-from-unr-gtlds There was an incident several years ago, where some gTLDs were moving between registries, and they were like "IDN is hard, so we're going to delete some domain names, it's fewer than 50 so who cares". And then Domain Incite wrote it up, and the new registry backend operator was suddenly like "NO WE
'The novelty now is the concentration of cases. '... a previous surge in 2024 when Ukraine's allies began allowing it to use Western-supplied weapons deeper inside Russia. 'But the attacks have continued. German media on Friday quoted a federal police report that lists more than 165 suspected sabotage cases nationwide this year, though they don't say who is responsible. 'They are usually Russian-s
Climate scepticism is common at the Reform party conference, but many also criticise Labour for failing to protect Britons from hot weather On the second day of the Reform UK conference, in the wake of an undercover investigation alleging party officials plotted to evade the law on foreign donations, the deputy leader, Richard Tice, addressed the audience. Within minutes, a banner appeared behind
The rough plan, which will probably be immediately amended upon contact with reality: Hook up the Pi with a TTL serial cable, Ethernet, and power it through a relay board. Build a minimal U-Boot that chainloads a full U-Boot over serial, flash that to the SD card Build a full U-Boot (with my patches) that sets up netconsole and loads a signed FIT image with kernel, initramfs, and maybe devicetree
Here’s a wee story of how I photographed this urban bird of the Andes, which happened to be a lifer for me (i.e., the first time that I have ever observed this species.) I was sitting in my office, on an audio call with a client. I saw this bird through my window hopping around frantically in the tree just outside. I grabbed my camera and I started taking photos, all while continuing to talk to th
I've just released my first ever MacOS app! It's something I've been needing for a while and decided to finally build it. I hope others find it useful as well. If you end up purchasing it, please do let me know as I'd love to hear any and all feedback that you might have! TopStart is live. For a while now I've been building a small Mac app quietly, and today it's done: https:// topstart.app is liv
There seems to be some interest, so I'm going to stream at ~4ish UK time on Twitch (sorry, don't have the energy to figure out a new platform today). Repeating the rough plan: Hook up the Pi with a TTL serial cable, Ethernet, and power it through a relay board. Build a minimal U-Boot that chainloads a full U-Boot over serial, flash that to the SD card Build a full U-Boot (with my patches) that set
;haiku You're right to ask. Here's the RCA: Root Cause: I violated my own guidelines repeatedly. 1. Made unverified claims instead of checking code - Said "face agreement test passes" when it was self-comparison (tautological) - Said "edge geometry validated" without actually testing both faces independently - Claimed "gap is rasterization artifact" without proof 2. Didn't read error feedback care
🚨🇧🇷 CBFS Academy CRM dataset allegedly offered for sale on a cybercrime forum, 43K+ leads claimed ⠀ CBFS Academy, a Brazilian football coaching and EdTech organization, is named in a cybercrime forum post where a threat actor claims to be selling live Supabase access to its CRM data. ⠀ The advertised dataset includes: ⠀ • 43,498 unique leads • 27,114 unique email addresses • 39,149 unique phone
TL;DR: I effectively have three different enumerations that need to be kept in lockstep. I'd like to turn them into one enumeration that can be used effectively for downcasting and rendering variant-specific data. My application supports two different RPG systems, and I'm adding a third one (and now suddenly feeling the pain of my architecture). So, use cases for my enumeration: - Whenever a chara
>The Reagan administration pulled right-wing media executives Rupert Murdoch and **Richard Mellon Scaife** into a CIA-organized “perception management” operation which aimed Cold War-style propaganda at the American people in the 1980s, according to declassified U.S. government records. >Although some records relating to Murdoch remain classified, several documents that have been released indicate
Apple’s iOS 27 is expected to include iPhone Handoff, a feature that lets users switch one phone number between two iPhones. The software release is expected later this month [theverge.com#1][google.com#1][bug.hr#1]. Apple briefly mentioned the feature during its WWDC keynote in June but did not explain how it would work [theverge.com#1][bug.hr#1]. Setup details surfaced after a MacRumors forum me
When I went off to college in 2001 my grandad bought me a map of Baton Rouge in a gas station on the drive there. Mapquest wasn’t a thing yet, or at any rate I didn’t yet have access to it. I knew how to read maps because he’d taught me while I was growing up. We had taken trips down to southern Louisiana to buy dogs, sell raccoons (don’t ask), and visit Avery Island, and he’d taught me about the
I'm working on an article but I've been super busy lately and distracted by other things, so why not keep the distractions going with a little tiktok ghost video. https://www. tiktok.com/@ljtherapy/video/76 80674182950112542 I've mentioned before that you have to be real careful with interpreting video from surveillance camera (or in this case I believe game camera) sensors under infrared illumina
When you've lived a few decades and had a lot of relationships that didn't work out, plus one or two really important ones that ended up not like you thought and more painful than you ever thought could happen to you , it's hard to lean into love and romance. Hard to just go with it 100%. Like, really hard. Maybe you can get to 80% or 90% or, after a few years, maybe 95%. But you can't just let go
Promise Backlund (@eve_wasframed) is a former worship leader turned counter-apologist. I first heard of her on Seth Andrews’s podcast. She wrote a book called Gospel of Lies (the title sounds harsher than it actually reads). There’s one chapter where she discusses the impact that music, architecture, mood lighting, and large gatherings can have in producing feelings of transcendence and euphoria f
Even more powerful: creating languages. Creating a language and its compiler/interpreter allows you to separate denotational semantics (what the programmer / # llm sees and reasons about) from its operational semantics (what actually happens). That means that by creating a language that stays close to the prompts a user would input (or… how humans communicate about the code), you get all the benef
[Digger/CH] # opencanary analysis for yesterday Summary: 🔄 Total Connection Attempts: 26303 👤 Unique Usernames: 866 🔑 Distinct Passwords: 1768 🌐 Unique Attacker IPs: 410 Port Popularity (Port / Count): 🔐 SSH: 11067 🖥️ RDP: 8103 🖥️ VNC: 5226 💻 Telnet: 1442 🗃️ MSSQL: 353 🔴 REDIS: 72 📡 Synology DSM: 17 🐬 MySQL: 17 📂 FTP: 3 🗄️ SMB: 3 Top 10 Usernames (Username / Count): 👤 root: 2092 👤
I recently installed Linux (Pop!_OS) on two devices that were collecting dust. A 13" Macbook Pro (Early 2015), and a Surface Pro 8 (2021?). The Macbook was EOL on OS updates and the Surface, well...had Windows (Home) on it. Both went surprisingly smooth. The Macbook basically worked out of the box, with the exception of needing to install brightnessctl and setup some keybinds for the keyboard back
Schools and councils express financial concerns as EHCP applications surge before potential change in provision ‘Powerful example’: minister looks to Essex school for blueprint to transform Send provision Schools and councils across England are expecting the proportion of children with special needs plans to double to one in 10, with families rushing to secure support before ministers bring in cha
Sporadic laser fire pierced the air of old San Francisco. The government of Cascadia had authorized civilian shootdowns of anything in the sky not bearing the flag of the Republic. With air travel too risky, a lone figure who’d made the almost as treacherous land crossing from the old country walked through the rubble of old Market Street. Sparking electrical boxes punctuating the wanderers footst
More than 153 million US and Canadian driver’s licenses, as well as other identity documents, have reportedly become available for purchase on the dark web for a limited time. According to cybersecurity journalist Brian Krebs, the service was called Nexus, and although it’s no longer available at the time of writing, it claimed to have possessed 153 million driver’s licenses, 10 million ID cards,
German AfD Leader Alice Weidel on Russia-Ukraine: The Alternative for Germany has a very clear position: we are always open to diplomatic channels. And we have always criticized the fact that, especially when we talk about Ukraine, we need an open channel to Russia in order to reach an understanding and achieve a peaceful solution. We have also always said that we need good relations not only with
British, Norwegian and U.S. forces deployed to the remote Norwegian island of Jan Mayen for Operation Atlantic City from August 26 to September 4 as part of NATO’s Arctic Sentry activity in the High North [ukdefencejournal.org.uk#1]. The exercise included Royal Marines, RAF personnel, the Royal Norwegian Navy, the Norwegian Home Guard and the U.S. Marine Corps. Allied forces trained in an environm
I've tried to fix this repeatedly over the years, and built a number of workarounds that mostly just pushed the problem from one place to another. Performance was less bad, that's about it. I'd spend a couple hours looking at it and eventually run out of information. Then I'd have to spend a half hour trying to figure out if (say) MacOS's Mellanox Ethernet driver supports LRO out of the box (yes),
I really need to write a contributing document (or at least statement) for musl to publish on the website. Basic principles: 1. Do not just send unsolicited patches except for trivial bugfixes. 2. If you want something not already slated for inclusion, or that's theoretically supposed to be but hasn't made progress, reach out first to explain what and why and find out if there are blockers or what
my wife was off inside still, and all the young children were inside and out of earshot of all of this, thank goodness, but my wife got out there about the time he had been sent off by his own wife, hes talking his shit and all this on his way to the fence, lol, and i was watching him to make sure i didnt need to act still and was so focused on him i couldnt feel her pulling my hand in retrospect
Not a # TIL (Today I Learned) about Linux, but a common misunderstanding I see in every Arch installation guide that isn't the ArchWiki: People want to enable TRIM to conserve their SSD lifespan, and thus they execute the following command: ``` systemctl enable --now fstrim.timer ``` That command by itself does absolutely nothing. All it does is create a scheduled job to trim (that is, discard unu
arXiv:2609.04382v1 Announce Type: new Abstract: We present a systems-security case study of a two-node split-LLM training system whose privacy evaluation passed while leaving an observable channel untested. The Trusted Local Node (TLN) sends protected activations to the Untrusted Cloud Node (UCN), the UCN returns its output, and TLN, holding the private loss, returns the output gradient. The frame
arXiv:2609.04410v1 Announce Type: new Abstract: Large language models can insert workplace details into phishing pretexts at low cost, but those details may either support or undermine a message's credibility. We recruited 180 U.S. working adults to evaluate simulated, AI-generated phishing emails in a disclosed survey. The emails used four cumulative levels of information: workplace (Level 1); re
arXiv:2609.04474v1 Announce Type: new Abstract: Phishing remains one of the most persistent cyber threats, yet publicly shareable datasets for studying phishing in realistic enterprise email settings remain limited. To address this gap, we introduce a synthetic enterprise phishing email dataset built around a fictitious organization, Nebulon. The dataset spans a broad set of workplace communicatio
arXiv:2609.04475v1 Announce Type: new Abstract: Federated unlearning aims to remove a client's data from a shared model without retraining from scratch. Some efficient systems make deletion exact by storing compact, additive summaries of the training features and broadcasting an updated linear classifier after every accepted change. We show that these broadcasts can also reveal the hidden summarie
arXiv:2609.04533v1 Announce Type: new Abstract: Prompt injection is widely recognized as a major security threat to AI agents that interact with untrusted external data, such as websites, documents, and emails. Prior work has shown that, in the text domain, black-box prompt injection can achieve near-perfect attack success rates (ASRs). In the image domain, however, existing visual prompt injectio
arXiv:2609.04769v1 Announce Type: new Abstract: Fully Homomorphic Encryption (FHE) enables computation on encrypted data, preserving privacy throughout analysis. While its privacy is very strong, FHE is much slower to execute than the original computation. In particular, due to the recent success in accelerating its compute, the performance bottleneck shifts to the memory, especially considering t
arXiv:2609.04785v1 Announce Type: new Abstract: Electromagnetic (EM) side-channel leakage and injection are typically treated as distinct physical phenomena, threatening data confidentiality and integrity respectively. This work investigates how EM injection can be used to amplify side-channel leakage that is otherwise infeasible. We introduce a novel framework for Injection-Induced EM Side Channe
arXiv:2609.04820v1 Announce Type: new Abstract: Ransomware detection and family attribution require analysis of different modalities because it can use packing, obfuscation, process manipulation and runtime evasion techniques. However, conventional multimodal usually uses all available modalities for every sample resulting in unnecessary computational cost and increased latency. In this paper, we
arXiv:2609.04875v1 Announce Type: new Abstract: Long-running LLM agents are stateful: beyond the transcript they accrete compressed summaries, plaintext memory, pending tool plans, and, under every serving API, a KV cache. Yet today's "forget" operations delete a plaintext memory record and stop, leaving every artifact derived from the revoked information intact. We formalize execution-state unlea
arXiv:2609.04878v1 Announce Type: new Abstract: Fraudulent messages sent via Short Message Service (SMS) are increasingly obfuscated to evade cost-conscious classifiers in production systems. In Chinese SMS, attackers can exploit a wide range of carefully crafted obfuscation strategies to hide risk-bearing phrases while preserving human readability, making direct classification brittle under real-
arXiv:2609.04970v1 Announce Type: new Abstract: Linguistic steganography enables covert communication through natural language. Existing methods heavily rely on token-level operations and struggle to maintain reliability under word- and sentence-level textual perturbations. Moreover, variable-length coding-based schemes are highly susceptible to bit-slippage under minor disturbances, as perturbati
arXiv:2609.05013v1 Announce Type: new Abstract: Multiple-input multiple-output (MIMO) technology is fundamental to modern wireless communication. Physical layer security seeks to protect transmitted information by exploiting properties of the noisy communication channel. Dean and Goldsmith proposed a polynomial time reduction from lattice problems to MIMO decoding by adapting Regev's reduction for
arXiv:2609.05092v1 Announce Type: new Abstract: Replacing classical entropy with QRNG output does not change honest Bitcoin PoW success probability when candidate headers remain distinct. The original contribution of this paper is a reproducible benchmark that locates and measures the operational value of quantum entropy in hybrid quantum-classical mining infrastructure through two scheduler-level
arXiv:2609.05117v1 Announce Type: new Abstract: Current LLM safety benchmarks largely rely on binary metrics, overlooking how models respond to harmful prompts with varying threat implicitness. We introduce TIER, a Threat Implicitness Benchmark for behavioral safety evaluation of LLMs. TIER covers four risk domains and four threat levels, from explicit harmful requests to sophisticated jailbreaks.
arXiv:2609.05119v1 Announce Type: new Abstract: Website fingerprinting (WF) attacks can infer which webpage a user visits from encrypted HTTPS traffic alone, compromising privacy even without decryption. WF defenses commonly shape traffic through noise, padding, delays, or flow splitting, yet they are most often studied from the perspective of encapsulating protocols like Tor or VPN rather than at
arXiv:2609.05165v1 Announce Type: new Abstract: Despite its introduction more than a quarter century ago, Conformal Prediction (CP) has seen surprisingly few applications to the cyber security world thus far. In particular, we observe that, while CP has been employed as a defensive measure in many recent works, its use for carrying out attacks (i.e., for offensive security) is hard to trace in the
arXiv:2609.05269v1 Announce Type: new Abstract: LLM agent systems increasingly combine provenance tracking, authorization, policy enforcement, protocol adapters, and execution controls. However, individually correct security mechanisms do not necessarily compose into an end-to-end secure system: security-critical context may be dropped, widened, rebound, or reinterpreted as actions cross component
arXiv:2609.05335v1 Announce Type: new Abstract: Public vulnerability databases collect rich information about known software flaws, including their weakness types, affected components, and related patches. Fixing commits provide the exact code changes that removed these flaws. While these records capture why the original code was unsafe, they are documented mainly for human inspection rather than
arXiv:2609.05370v1 Announce Type: new Abstract: Decompilation recovers high-level source from compiled machine code and serves as a foundation for security tasks such as vulnerability detection and malware analysis. Traditional decompilers like Ghidra and Hex-Rays expose whatever they cannot resolve as visible placeholders and often emit pseudocode that will not compile or execute; LLM-based decom
arXiv:2609.05380v1 Announce Type: new Abstract: Ensuring security of software supply chains (SSC) is indispensable in today's world of modern software practices. SBOM (tools) have been introduced as relevant building blocks to ensure the transparency of SSCs. However they have serious limitations in practices as their vulnerability detection and interpretation capacity is not sufficient to explain
arXiv:2609.04495v1 Announce Type: cross Abstract: We formulate indirect prompt injection as a test-time search over a task-dependent attack surface induced by the environment, user task, and injection task. To operationalize this formulation, we introduce an agentic attacker with a dedicated search harness that performs environment reconnaissance, structured reasoning over attack strategies, and a
arXiv:2609.04815v1 Announce Type: cross Abstract: Detecting orchestrated cyberattack campaigns that span multiple organizations traditionally requires sharing sensitive telemetry and threat intelligence across institutional boundaries and country borders, a barrier that Federated Learning removes by training shared threat detectors directly on local data. We propose FedIoC, a modular framework in
arXiv:2609.05340v1 Announce Type: cross Abstract: Agent based systems are increasingly deployed in information critical systems including healthcare management systems, and smart grids. In this paper, we consider a multi-agent system where each agent has a latent goal that needs to be kept hidden from observing adversaries. More specifically, this paper studies privacy-preserving consensus in netw
arXiv:2605.29354v2 Announce Type: replace Abstract: LLM-powered coding agents increasingly participate in software development workflows by generating code, selecting dependencies, and producing package installation commands. This creates a new software supply chain risk: when an agent hallucinates a non-existent package, an attacker may register the hallucinated name and later compromise users wh
arXiv:2606.03090v3 Announce Type: replace Abstract: The emergence of large language models (LLMs) has significantly accelerated recent research on LLM-based automatic grading (AG) systems. Benefiting from the strong instruction-following capabilities and broad prior knowledge of LLMs, educators can deploy AG systems across diverse tasks using only natural language rubrics while achieving satisfact
arXiv:2606.26390v2 Announce Type: replace Abstract: While several web actors have been trying to reduce web tracking for years, it remains unclear how to achieve both desirable levels of utility and privacy. In 2019, Google launched the Privacy Sandbox initiative to balance that trade-off and find privacy alternatives to common use cases such as advertising. Yet, in late 2025, Google canceled the
arXiv:2607.13596v2 Announce Type: replace Abstract: When cast as the protector of a vulnerable user yet given no explicit capability boundary, a large language model (LLM) may respond not by acknowledging its limits but by claiming to have taken, or to be taking, a real-world protective action it cannot perform, such as contacting emergency services or administering care. We term this phenomenon P
arXiv:2506.21114v5 Announce Type: replace-cross Abstract: To cater to the needs of fast verification for mathematical proofs, we describe a method to encode formal sentences in $2 \times 2$ - matrices over multivariate polynomials with integer coefficients. This correspondence is homomorphic: usual proof-steps like modus-ponens or variable substitution in terms and formulae become operations with
arXiv:2603.18077v4 Announce Type: replace-cross Abstract: Code smoothing is a phenomenon in which an error distribution makes a code statistically close to the uniform distribution over the ambient space. This closeness is measured by the total variation distance. Recently, Debris-Alazard et al.\ introduced a smoothing bound, which is an upper bound on this total variation distance. Although the s
arXiv:2609.04496v1 Announce Type: new Abstract: Modern portfolio theory identifies diversification as the primary tool for risk reduction. However, under model uncertainty, this cornerstone may no longer remain optimal. This paper investigates the tension between portfolio diversification and concentration under dependence uncertainty. In the absence of model uncertainty, we employ the framework o
arXiv:2609.04569v1 Announce Type: new Abstract: Herein, we propose a quantum circuit learning framework for modeling the realized volatility (RV) of Bitcoin and investigate the statistical properties of the predicted time series through multifractal analysis. Unlike conventional GARCH-type models, which require a pre-specified functional form for the volatility process, a parameterized quantum cir
arXiv:2609.05115v1 Announce Type: new Abstract: Creativity researchers often distinguish between two stages of the creative process: generation versus selection. While much is known about the psychology of idea generation (e.g., the factors that lead to a greater number of novel and useful ideas), less is understood about the nature of selection, or how generation and selection interact. Here we i
arXiv:2609.05162v1 Announce Type: new Abstract: We propose a new model of expected stock returns that incorporates quantity information from market trading activities into the factor pricing framework. We posit that the expected return of a stock is determined by not only its factor risk exposures (beta) but also the factor's quantity fluctuations (q) induced by trading flows, and hence term the m
arXiv:2609.04420v1 Announce Type: cross Abstract: Let a finite population of n labelled examples carry a class-weighted loss, with pi*n in a rare positive class weighted by N0/N1. We study estimation of total risk from a subsample K << n under designs allocating K0 and K1 draws to the two strata. We derive the exact finite-population variance of the weighted risk estimator under class-conditional
arXiv:2609.04712v1 Announce Type: cross Abstract: Transaction costs can make or break a trading strategy, particularly in relative-value trading of commodity and macro markets, where edges are a few basis points. Price impact is a central component of transaction cost. Price impact models usually include self-impact (a trade in a contract moves that contract's price) but omit two well-documented e
arXiv:1302.4676v4 Announce Type: replace Abstract: The multilevel Monte Carlo path simulation method introduced by Giles ({\it Operations Research}, 56(3):607-617, 2008) exploits strong convergence properties to improve the computational complexity by combining simulations with different levels of resolution. In this paper we analyse its efficiency when using the Milstein discretisation; this has
arXiv:2310.00553v4 Announce Type: replace Abstract: Financial institutions often cannot replicate long-dated liabilities with available bonds, especially when leverage and collateral constraints bind. We characterize the feasible portfolio that minimizes, to first order, the worst equity loss over a prespecified set of yield curve changes. The solution accommodates general deterministic liabilitie
arXiv:2510.01211v2 Announce Type: replace Abstract: We provide series expansions for the tempered stable densities and for the price of European-style contracts in the exponential L\'evy model driven by the tempered stable process. These formulas recover several popular option pricing models, and become particularly simple in some specific cases such as bilateral Gamma process and one-sided TS pro
arXiv:2607.00245v2 Announce Type: replace Abstract: Autonomous artificial intelligence (AI) agents are beginning to occupy a position between analytical tools and transacting counterparties. They can interpret goals, call external tools, negotiate with other agents, access data and computation, and in some settings initiate payments or blockchain transactions. This development creates a distinct p
arXiv:2607.09951v2 Announce Type: replace Abstract: This paper develops a model of maritime chokepoint closures in which interrupting a shipping passage produces losses that are not measured, or even bounded, by the value of the trade that transits it. The losses stem from disruptions to the flow of intermediate inputs that are complementary in downstream production. Re-matching displaced trade on
Mommy wished on her shed whiskers: canned food every day for her, plus a tiny side wish for a bigger house and more bags. She knows about the side wish. The whiskers now live in their own tiny vase under daily supervision. The wish came true. The apology is still pending. For context: In some countries, a cat's shed whisker found by chance is a wish-granting lucky charm. Credit: Dry-Reveal4114 ![]
China is actively researching humanoid robots for military use. Reuters reviewed 100+ procurement records, patents, and military papers showing the PLA planning for eventual battlefield deployment. A 2025 military university paper modeled humanoid robots in an urban assault force clearing buildings alongside troops. The PLA's newspaper called for accelerating transfer of robot tech from labs to tr
Torrential rain causes widespread damage; while north-west South Africa hit by thunderstorms and central-southern US faces oppressive heat Typhoon Saudel brought days of torrential rain to the Zhejiang, Jiangxi, Fujian and Guangdong provinces in south-east China last week. The city of Putian, in Fujian province, recorded 416.5mm of rain in a 24-hour period between Thursday and Friday, with surroun
US-owned dating app allegedly allowed ad firms access to private information in breach of UK law Business live – latest updates Grindr will pay £26m to settle a lawsuit brought by thousands of users in the UK who allege that the US dating app shared highly sensitive personal information, including in some cases their HIV status, with advertising companies. The settlement brings to an end a two-yea
This affects the Fediverse. This affects the Internet. This is the equivalent of the "rubber hose method" of decryption. All technical solutions (read: all technology) is subject to politics. To political power. To political interactions. You cannot rely on purely technical approaches. You have to have social approaches, too. Like. Actual policy and procedure geared around social and political gro
In yet another case of churches´anti-AfD standards for parishioners and employees of church-affiliated organizations, a Protestant church near Oldenburg asked a woman to resign from volunteer roles because her partner is an AfD candidate in a local race. www.nwzonline.de/oldenburg-kr... Marlene P. aus Hundsmühlen engagierte sich jahrelang ehrenamtlich in der evangelischen Kirchengemeinde. Weil ihr
cross-posted from: https://lemmy.world/post/51634640 > A new Gamers Nexus investigation found a retail LG TV actively scanning the local network for phones, laptops, smartwatches and other connected devices, while also capturing microphone audio when the screen appeared to be in standby. The investigation found voice data being stored locally even after the TV was disconnected from the network, wi
An unnamed member of the SF Opera Chorus: Whenever the subject of “financial challenges” appear, it is the artists (instrumental, vocal, backstage, etc.) who are expected to make the sacrifices. These are the people who make the product which these organizations are selling. Workers are expected to make do with less, still perform and do their best to inspire and uplift the people who are able to
SEPT 7th Today we have Bolivian cucumber achocha West India Burr Gherkin Litchi tomatoes Pickling cucumber Tomatillo Ground cherries Peas Pumpkins: Neon pumpkin (sugar), Kabucha (Japanese pumpkin) orange Winter squash: Georgia candy roaster, Red kuri, baked potato, Gete Okosomin SEPT 7thwe are open 10am to 8pm9632 69a st NWSelf-ServiceCash or e-transferToday we haveBolivian cucumber achochaWest In
this outlet adds: peppers, mirchi, pattypan, yellow, green, white, custard, trombocino
things I believe: - # llms represent a fundamental change for the better to how we build software - "agentic" coding is nonsense - # llms are not "intelligent" - # llms are statistical models of language for increasingly alien values of "language" - language running on its own like a wind-up toy can do a lot of things that only "intelligent" people used to be able to do - we should pause model dev
Based on what I've seen on Youtube, the speedrunner meta calls for taking the Closet, in the position shown above, as your first play: there's RNG for what you find in the closet, and if you don't get something good ... well, you haven't wasted much time, just restart. I'm not sure it's the best call for someone just trying to land the achievement once. My opening here was to pick Closet / Parlor
RE: https:// mastodon.online/@rozie/1172242 44824728530 Eteryu stara się przedstawić swoje tezy jako wynik rzetelnej analizy, jego tekst zawiera elementy dezinformacji, wybiórcze traktowanie faktów oraz jednostronne podejście do omawianych kwestii. Jak na moje oko, Baza wiedzy Eteryu space – jest jedynie kopią, niestety mocno okrojoną i uproszczoną. Szczegóły w dalszej częśći tekstu, wraz ze zródł
Hey HN! Ilya and Nikita here. We’re building HomeCat, a tool that lets you design a backyard office space - https://myhomecat.com We started out by helping people design sheds, and quickly realized that many people actually want to build a backyard office. There wasn't really an easy way to do this, so we thought we'd take on the challenge. Right now, you can design a backyard shed and our tool wi
I wrote this 11 YEARS ago. And it only got more true. "Similarly, the technologies of personal data capture interlace consumerism, technology, and citizenship. Under this framework, the implications of resistance or “opting out” become more severe. [17] Using Tor to evade consumption-based surveillance flags particular national security watch-lists; extensive withdrawals of cash from various ATMs
RE: https:// hachyderm.io/@davidculley/1171 45285699647062 Today I was mentioned by someone who is very likely a genuine victim of genicide but who isn't yet # GazaVerified by @ gazaverified (because they created their account only today). I was one of many people who were mass-mentioned by this unsolicited message. I wrote to the person behind that account that I'm sorry for the suffering they ar
Move in at the college for 2030 class was the other week. I was in the Square today and it was totally busy. People everywhere, shops busy, it was like the old Square again. However... Your children. Fucking undergrads. A girl stopped me and asked me if I knew where the Verizon store was. I said "Yes, I'm pretty sure it's this way" and I pointed. Her reply was 'okay I'll keep asking people.' To wh
# MakeShitMonday , in which I return to a # crochet project that's been languishing for a couple years, inspired by @ consumablejoy and her # FinishOrFrogAlong adventure... Back in April of 2024 I made a crochet rat for @ mbroome 's niece, due to a rather hilarious mixup where I offered to make her a wrap, she thought I said "rat", and suddenly I had a [new amigurumi project]( https://www. withlov
The Federal Communications Commission (Commission or FCC) issues a Notice of Proposed Rulemaking proposes to expand opportunities for direct-to-device (D2D) communications by permitting certain unlicensed devices to communicate with satellites on a non-interference basis. Building on recent market growth and new industry investment in D2D technologies, the NPRM explores adding satellite allocation
Shinjuku ward targets accommodation in residential areas and near cultural sites as locals complain visitors disposing of rubbish improperly and being too noisy A neighbourhood in Tokyo popular with foreign tourists will ban visitors from using thousands of short-term rental accommodations after a slew of complaints about behaviour. In the latest sign that Japan’s patience with mass tourism is wea
Via Gwynne Dyer, who I've followed for many years and long ago correctly predicted our complete failure to reduce CO₂ emissions, on the consequences of that failure. Geoengineering has been, for good reasons, very unpopular among many, but I don't see a way to avoid his conclusion below. "Will [the 2028 El Niño] trigger another big ‘non-linear’ jump in the planet’s temperature? Nobody knows, actua
Campaigning on health care used to mean talking about coverage or costs . Now, in the run-up to the midterms, it's just as likely to focus on "corruption." The big picture: Candidates are accusing primary or general election opponents of being in the pockets of drugmakers, insurers and other powerful health interests, to the detriment of voters. It's a strategy that plays into affordability concer
The Nazis didn't start with extermination camps. They built their extermination camps (most infamously: Auschwitz) in 1942. From 1933 until 1939 they imprisoned Jews (after stealing all their stuff) to force them to "voluntarily" emigrate. The death rate in concentration camps such as Dachau (due to torture and forced labor or arbitrary murders out of revenge) was at 4%. "Voluntary emigration", de
Interesting research: “ Stealing Reasoning Traces from Proprietary LLM APIs “: Abstract: Leading large language model providers now conceal their models’ step-by-step reasoning, or chain-of-thought, to protect intellectual property and limit information leakage. Rather than storing these traces server-side, providers return them to the client as blocks of encrypted text, which the client passes ba
Brent oil surged toward $100 a barrel after Saudi Arabia halted some energy operations near its border with Yemen due to attacks, highlighting risks to supply from the region. Novartis shares plunge as much as 10%, marking the worst day since 2020, after its del-desiran treatment for a rare muscle-wasting disease failed to meet the primary endpoint in a phase 3 trial. Vontobel said this is another
It’s a fucking nightmare to live in the US today, how the fuck is it that we’re implicitly champions of the industrial world but we fucking failed at having functional homes for people in it? Seriously, you wanna go across the street you need a fucking car to go out your bullshit apartment. You wanna go spend time with friends? You’re a loser if you don’t have wheels. You’re trying to go to that i
BREAKING: Twelve countries issued a joint statement announcing national bans on trade in goods from illegal Israeli settlements: 🇫🇷 France 🇬🇧 United Kingdom 🇨🇦 Canada 🇩🇰 Denmark 🇪🇸 Spain 🇫🇮 Finland 🇮🇪 Ireland 🇮🇸 Iceland 🇳🇴 Norway 🇵🇱 Poland 🇵🇹 Portugal 🇸🇪 Sweden Macron, UK PM Burnham, and Canadian PM Carney agreed the situation is deteriorating with "unprecedented" settler v
OK: this is good @ paulkrugman.substack.com "Paul Krugman: Where's the Beef (Coming From)?" "So you may have seen that last month Trump announced, given that we have high beef prices, that he was going to allow the tariff-free import of 300,000 tons of ground beef. Rather oddly, he didn’t say from where and wouldn’t say from where for a while. And then eventually said, well, from Argentina, Brazil
Confirmed accurate and fresh (filed this week). Here's the tweet: --- 📸🕶️ Meta is being sued in a new nationwide class action alleging it quietly mined Facebook and Instagram photos to build NameTag, an unreleased facial recognition system meant for its smart glasses, without telling users. The lawsuit also claims those same photos were used to train Meta's generative AI, again without consent.
Scientists used NASA’s Hubble Space Telescope and James Webb Space Telescope together for the first time to study trans-Neptunian objects, small icy bodies that orbit the Sun beyond Neptune [nasa.gov#1][phys.org#1][perfil.com#1]. The study examined 27 newly detected objects that are unusually small and faint, and found fewer small TNOs than researchers expected [phys.org#1][perfil.com#1][wp.pl#1][
🚨🇨🇴 Unidad Central del Valle del Cauca allegedly breached, 4.1K+ individuals affected ⠀ Unidad Central del Valle del Cauca (UCEVA), a public higher education institution in Colombia, is named in a cybercrime forum post where a threat actor claims to have obtained student PII, staff account information, internal documents and data from exposed university systems. ⠀ Claimed exposure ⠀ • 4,090 stu
Attackers sent 2.37 million messages a day without hiding anything. They put a character you cannot see in the middle of the words email gateways scan for. "Funding" becomes "fun" plus an invisible character plus "ding". Your users read it normally. Your keyword rule stops matching entirely. Microsoft and Fortra documented the campaign this month. The Framework now does this at send time against 1
I have two previously torn meniscuses and a torn patellar tendon. I don't run a ton (my workout program does agility training for cardio), but I can just decide to go run a sub 30m 5k with no problems. It definitely wasn't always this way. I used to have to tap after 10 minutes due to the knee pain. Until one day I thought, "Maybe I'm doing something wrong," so I did some research. Turns out I **w
"A proposed budget cut would reduce Medicaid reimbursement rates by 38% for five key categories of behavioral health providers: clinical social workers, psychologists, marriage and family therapists, residential treatment facilities, and alcohol and drug treatment centers." -- @ sltrib # Utah already throws piles of money at an abusive # Synanon model in # TOSA (The Other Side Academy) and at poli
Rethinking On-Policy Distillation of Large Language Models II: One Training Example On-policy distillation (OPD) combines student-generated rollouts with dense token-level supervision from a teacher. Existing work has mainly studied its algorithmic behavior, leaving the role of training data unclear. We examine this role at the data-minimal limit… arXiv — Computers & Society + AI · cs.CY, cs.AI up
I have been test driving Astra/5.6 for designing and implementing relatively complex Storage project in Openshift/Kubernetes land. Here is my findings so far when it comes to API design: - Astra produced APIs which are not self-consistent. Think a field's default value is incorrect when a higher level value is set to something else (why is it default then?) When it came to implementation: - I had
SSH is supposed to keep attackers out, not let them in. But router vendor MikroTik has patched a bug that did just that… CERT Polska found a bunch of bugs in MikroTik’s SSH server, which seems to be a home-made or home-modded version, rather than a regular build of trusted source code such as OpenSSH. Two of these bugs could be chained together to let anyone in by default, and then promote them to
I particularly like breaking down the javelins. They are like intricate little puzzles and there are a bunch of different ways they can be weird in ways you have to work around. Like one I did recently had the rear section blocked by a giant fuel tank, so there was literally no way in without disassembling the front part of the ship. I _did_ that and freed it from the front part, but in doing so i
‼️ FortiSandbox Vulnerability Allows Attackers to Access Sensitive Information via Crafted HTTP Requests Source: https:// cybersecuritynews.com/fortisan dbox-vulnerability-access-sensitive-data/ The flaw, tracked as CVE-2026-26084, stems from improper access control in the graphical user interface component that FortiSandbox, FortiSandbox Cloud, and FortiSandbox PaaS all share, and it has been ass
arXiv:2609.05627v1 Announce Type: new Abstract: Contactless payment from a phone has become routine in many markets, but the convenience is unevenly distributed. Apple currently lists Kazakhstan among supported Apple Pay markets, but not Uzbekistan, Kyrgyzstan, Tajikistan or Turkmenistan. In Uzbekistan, HUMO and UZCARD are the two national interbank retail card systems, and both include contactles
arXiv:2609.05901v1 Announce Type: new Abstract: Autonomous LLM agents can turn untrusted content into effectful actions such as payments and permission changes. If the same process interprets this content and controls a reusable signing credential, prompt injection can cross the judgment boundary and reach execution authority. We present KITA, a review-to-authorization architecture that keeps the
arXiv:2609.05903v1 Announce Type: new Abstract: Large Language Model (LLM) agents are turning language into real-world effects, making safety necessary against both indirect prompt injections and direct harmful requests. System-level safety harnesses add an enforcement layer beyond model-level defenses, but existing harnesses are usually designed once by experts and applied across heterogeneous mo
arXiv:2609.05911v1 Announce Type: new Abstract: Long-horizon LLM agents interact with untrusted content, persistent memory, external state, and sensitive tools. Existing analyses often characterize attacks by the number of execution steps between malicious input and a downstream action. We show that temporal remoteness can overstate security separation in stateful agents. We introduce a provenance
arXiv:2609.05920v1 Announce Type: new Abstract: Agent Skills combine instructions with files, packages, tools, models, and services, so operational identity can exceed a signed directory. Recursive or lazy dependencies may change while root-level evidence remains valid, and different surfaces may reach the same durable effect. We present ClosureBound, a reference monitor that prevents authorizatio
arXiv:2609.06012v1 Announce Type: new Abstract: Machine learning-based network intrusion detection systems (ML-based NIDS) are vulnerable to adversarial evasion, where malicious samples are perturbed to evade detection and be misclassified as benign. Despite growing research on adversarial attacks and defenses for ML-based NIDS, comparative evaluations of multiple attack types, detection models, a
arXiv:2609.06027v1 Announce Type: new Abstract: Search-augmented LLM agents are increasingly used for consumer decisions, making them vulnerable to Generative Engine Optimization (GEO) poisoning. Existing benchmarks largely measure whether manipulated content is retrieved or endorsed, but do not track whether an agent verifies suspicious evidence, revises adopted claims, or recovers before produci
arXiv:2609.06103v1 Announce Type: new Abstract: This research studies how Technical Support Scams (TSS) are being countered by a uniquely dedicated community of volunteer counter-fraud operatives. Using a careful subject selection strategy, we interviewed 17 individuals who actively engage in TSS scambaiting activities in order to obtain insight into their motivations, the operational methods of t
arXiv:2609.06140v1 Announce Type: new Abstract: Agents can turn shared infrastructure into a channel for coordinated intrusion. The Hugging Face incident and a separate public-wiki investigation show why a security assessment may need evidence from several executions and the artifacts they leave behind. We argue that the operational unit of defence should be a revisable coordination episode linkin
arXiv:2609.06149v1 Announce Type: new Abstract: Script-based malware remains a prevalent attack technique. These scripts often contain indicators of compromise (IOCs) that provide actionable threat intelligence. However, statically recovering such indicators is challenging, as relevant values may be dispersed or transformed within code. Although large language models (LLMs) have shown promise in s
arXiv:2609.06271v1 Announce Type: new Abstract: Some low-cost Internet of Things (IoT) sensor deployments lack device-level source authentication, leaving them vulnerable to impersonation or injected sensor readings. We present a lightweight approach to sensor impersonation detection in a small proof-of-concept study. We formulate detection as a sequence-prediction problem. A model with three LSTM
arXiv:2609.06306v1 Announce Type: new Abstract: Programmable Logic Controllers (PLCs) increasingly run multiple applications alongside the main control program, with shared access to PLC variables. Yet, Industrial Control System (ICS) defenses primarily detect malicious updates by checking whether variable values violate expected bounds, without considering which application performed the update.
arXiv:2609.06322v1 Announce Type: new Abstract: The Side-Channel Existence Theorem proves $I(X;Y)>0$ in the binary, undefended setting, but is confined to pairwise arguments and ignores active defenses. We extend it to $k$ classes via the per-class decomposition $I(X;Y)=\sum_i\pi_i D_{\mathrm{KL}}(P_{Y|i}\|P_Y)$, with defense cost modelled by per-class Wasserstein-1 constraints $\sup_x W_1(Q_x^D,P
arXiv:2609.06330v1 Announce Type: new Abstract: Large language models are often instruction-tuned, specialized, quantized, or otherwise transformed, making fine-grained provenance difficult. In this paper, we introduce BReF, a training-free fingerprint that compares how probability distributions over four answer-option labels A/B/C/D move under controlled textual perturbations. For each pair of mo
arXiv:2609.06412v1 Announce Type: new Abstract: Hardware for lattice-based post-quantum cryptography spends a large share of its area on the number-theoretic transform (NTT), dominated by modular multipliers and twiddle storage. FoldNTT is a redesign of the released radix-2 CFNTT accelerator (TCHES 2022) for the Falcon / FN-DSA prime q = 12289 with one hardware multiplier per butterfly instead of
arXiv:2609.06539v1 Announce Type: new Abstract: Typing passwords on smartphones in public places exposes users to video-based side-channel attacks, in which an adversary records the typing session and reconstructs the entered password by analyzing finger movements. Prior video-based keystroke inference attacks have targeted free-form text on tablets, numeric PINs, pattern locks, and passwords unde
arXiv:2609.06540v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed in safety-critical applications, yet jailbreak attacks can conceal harmful intent through role-playing, fictional scenarios, or seemingly benign motivations. Existing inference-time defenses may miss disguised attacks or excessively refuse legitimate requests. We propose SRD-GUARD, a parameter-fr
arXiv:2609.06579v1 Announce Type: new Abstract: Machine learning methods, and especially neural networks, are now routinely used for malware detection in network traffic. Though very effective, systems based on such methods often (i) are purely data-driven, ignoring the substantial body of available knowledge about the tactics, techniques, and procedures (TTPs) possibly used, and, consequently (ii
arXiv:2609.06674v1 Announce Type: new Abstract: We present a new attack that reconstructs the text generated by locally hosted LLMs by observing CPU cache activity during detokenization. Unlike prior attacks that rely on deployment-specific assumptions, such as shared data memory, CPU offloading, or Mixture-of-Experts architectures, our approach targets the detokenizer, a component used in default
arXiv:2609.06708v1 Announce Type: new Abstract: Large language models are transforming many industries with their text generation abilities. However, their outputs can be easily tampered with, creating serious risks in critical areas such as food safety reporting. To protect the integrity and traceability of AI-generated content, this paper introduces ToSS (Token Oriented Repartitioning and Strate
arXiv:2609.06773v1 Announce Type: new Abstract: Mutation-based fuzzing is widely used to discover software vulnerabilities, but its randomness complicates rigorous evaluation and reliable bug detection. Prior work measures this variability empirically but lacks a theory with computable convergence and sample-complexity guarantees. We address both problems. First, we estimate robustness from indepe
arXiv:2609.06783v1 Announce Type: new Abstract: LLM agents operate in workflows where unsafe actions can have real consequences. Existing safety evaluations often reduce behavior to a single score, obscuring risk recognition, pre-action detection, and safe task completion when a safe solution exists. We introduce AURA-Eval, a framework combining controlled augmentation with granular diagnosis of b
arXiv:2609.06791v1 Announce Type: new Abstract: Register-Transfer Level (RTL) simulation is widely used to test hardware before it is fabricated. To allow testing for security related information flow properties, such as confidentiality and integrity, taint logic can be automatically added to the design to track how information flows through it. However, taint logic instrumented by the state-of-th
arXiv:2609.06835v1 Announce Type: new Abstract: Agentic AI systems execute complex tasks through long-horizon workflows of planning, tool use, and multi-agent coordination. Task failures in these systems often originate from a single step, such as an injected prompt or a flawed plan, and are then amplified through downstream dependencies as the corrupted step propagates across many subsequent agen
arXiv:2609.06855v1 Announce Type: new Abstract: The Open Radio Access Network (O-RAN) replaces vendor-locked RANs with a modular and interoperable architecture that fosters competition and accelerates innovation. With this openness comes increased complexity and a larger attack surface, making security a critical concern. Today, assessing O-RAN security requires manually cross-referencing dozens o
arXiv:2609.06972v1 Announce Type: new Abstract: LLM agents complete tasks by issuing sequences of tool calls, and every observation they read is a channel through which an indirect prompt injection can enter. A successful injection has a characteristic shape when the trajectory is read in order: a benign prefix gives way to actions that serve the attacker rather than the user. Existing benchmarks
arXiv:2609.06973v1 Announce Type: new Abstract: Electromagnetic signal injection attacks (ESIA) pose a growing threat to image sensors, which are increasingly used in different intelligent systems. By emitting electromagnetic interference, adversaries can manipulate pixel values, potentially misleading downstream artificial intelligence (AI) models and causing unsafe decisions in these systems. We
arXiv:2609.07022v1 Announce Type: new Abstract: Reasoning over privacy-constrained records requires combining structured attributes with evidence from free-text narratives. We introduce CIPHER (Cross-record Inference over Privacy-Hardened Evidence Records), a benchmark of expert-validated questions from consumer-finance, clinical, and law-enforcement records. The questions cover common tabular ope
arXiv:2609.07131v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly achieve long-horizon tasks by combining foundation models with explicit skills and implicit procedural knowledge acquired through execution. The resulting task-solving capabilities have become valuable proprietary assets, raising a new security question: can a substantially weaker attacker-controlled age
arXiv:2609.07614v1 Announce Type: new Abstract: Cybersecurity vulnerability information is distributed across numerous platforms and databases, making it difficult for researchers and practitioners to obtain a unified and structured understanding of existing threats. This is a critical issue in cybersecurity, where timely access to accurate vulnerability information directly impacts risk assessmen
arXiv:2609.07641v1 Announce Type: new Abstract: Providers of continuously produced, commercially valuable data -- sensor streams, telemetry, and other feeds sold as machine-learning training material -- cannot observe whether delivered data is actually used, while data that keeps flowing to inactive consumers enlarges the leakage surface without producing value. We present Attestream, a blockchain
arXiv:2609.07654v1 Announce Type: new Abstract: GSMA Remote SIM Provisioning (RSP) enables over-the-air delivery of eSIM profiles, but it exposes long-lived identifiers during profile ordering and download. In particular, stable device identifiers (e.g., EID), profile identifiers, and long-lived certificate material enable mobile operators and profile-delivery infrastructure to link provisioning e
arXiv:2609.07709v1 Announce Type: new Abstract: SSH is the standard protocol for secure remote administration of servers. At the transport layer, SSH uses the Binary Packet Protocol (BPP) for encrypted and authenticated communication. Above this, the SSH Connection Protocol multiplexes one or more logical channels over a single connection, supporting interactive shells, port forwarding, and relate
arXiv:2609.07783v1 Announce Type: new Abstract: Grid-connected solar distributed energy resources (DERs), such as solar inverters and monitoring platforms, have been deployed at unprecedented scale over the past few years, with global solar capacity more than doubling since 2022. To support monitoring and control, many of these systems are Internet-connected and configured by installers or end use
arXiv:2609.07849v1 Announce Type: new Abstract: Post-quantum protection is delivered to a peer, not declared in a file: whether a session is quantum-resistant is a relation between a server's configuration and the clients that reach it. We show that no single peer can soundly gate that relation. Shipped SSH clients are not ordered: two of their post-quantum capability classes are minimal and incom
arXiv:2609.07865v1 Announce Type: new Abstract: In recent years, smart contracts have become the backbone of decentralized applications (DApps), and off-chain systems such as bridges, wallets, and indexers rely heavily on event logs to track contract execution and state changes. However, the Ethereum Virtual Machine (EVM) does not validate or enforce event semantics, so logs can diverge from on-ch
arXiv:2609.07963v1 Announce Type: new Abstract: Traditional light clients rely on validators committing to the entire blockchain state at every block via a state commitment such as a Merkle tree, allowing clients to verify facts using short proofs. However, maintaining large and ever-growing state trees imposes a significant burden on validators and lies on the critical path of block production. A
arXiv:2609.08009v1 Announce Type: new Abstract: Otome games, a romance simulation genre primarily targeting female, have emerged as a major force in the global gaming market, attracting hundreds of millions of players and billions in revenue. Despite their popularity, otome game communities face pervasive online toxicity, which has been largely unexplored. In this work, we present the first large-
arXiv:2609.08040v1 Announce Type: new Abstract: The software supply chain has become an increasingly exposed attack surface because of its reliance on intricate yet fragile dependencies. Existing defenses such as GitHub Dependabot often raise many false alerts because their coarse-grained matching cannot determine whether a vulnerable dependency is actually exploitable. Security analysts typically
arXiv:2609.08093v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly employed for offensive cybersecurity tasks such as automated vulnerability discovery, reconnaissance, and penetration testing. This new capability also threatens one of the defender's most valuable tools: deception. Traditional honeypots rely on realism and obscurity to lure human or script-driven at
arXiv:2609.08213v1 Announce Type: new Abstract: Diffusion watermarking embeds verifiable signals into the generative process and commonly verifies them by recovering trajectory-dependent evidence, making the marks robust to conventional pixel-space distortions. Existing removal attacks either regenerate along deterministic trajectories, which often preserve the watermark-bearing latent structure,
arXiv:2609.08256v1 Announce Type: new Abstract: Automated red-team attacks and blue-team defenses for large language models (LLMs) are advancing quickly. However, attackers and defenders are built and tested in isolation, and the resulting scores are hard to trust. To tackle this, we present ACEA (Adversarial Co-Evolution Arena), a platform that connects a pluggable red-team adapter and a pluggabl
arXiv:2609.08306v1 Announce Type: new Abstract: We introduce HoneyRoute, an inference-serving layer that detects whether an incoming request is malicious and, if so, routes it to a dedicated honeypot model, shielding production while the adversary's interaction is continuously harvested for intelligence. Existing defenses embed traps inside model memory or rebuild deception at the protocol layer,
arXiv:2609.08476v1 Announce Type: new Abstract: Secure Aggregation (SA) is widely regarded as a strong defense against model-update leakage in Federated Learning (FL), as it reveals only aggregate results while hiding individual updates. In Decentralized Federated Learning (DFL), SA is commonly instantiated as local neighborhood aggregation, where each node obtains a weighted aggregate over its ne
arXiv:2609.08481v1 Announce Type: new Abstract: Agentic AI systems increasingly exchange messages, invoke tools, request approvals, hold structured decision sessions, and modify shared artifacts. Logs and anchors can make selected records tamper-evident, but they can also mislead if their evidentiary meaning is implicit: a hash does not establish semantic truth, a signature does not establish auth
arXiv:2609.08623v1 Announce Type: new Abstract: Network intrusion detection systems (NIDS) are critical for cybersecurity, safeguarding services and data from potential attacks. However, existing AI-based NIDS often assume static data distributions and fail to handle concept drift, leading to degraded performance and increased false positives in dynamic network environments. To address this issue,
arXiv:2609.08763v1 Announce Type: new Abstract: Federated global navigation satellite system (GNSS) monitoring distributes a proprietary classifier to partly trusted stations, any of which may leak its copy. ZK-Trace combines public identity marks, recipient-specific Tardos fingerprints, and zero-knowledge credential verification. The registry supports offline tracing without the leaker's cooperat
arXiv:2609.08765v1 Announce Type: new Abstract: Large language model (LLM) benchmarks are often treated as fixed datasets with stable scores, yet their outcomes depend on configurable evaluation pipelines. We audit eight cybersecurity benchmarks across 10 proprietary, open-weight, and cybersecurity-specialized LLMs. By modeling benchmarks as measurement pipelines, we identify 15 systematic failure
arXiv:2609.08871v1 Announce Type: new Abstract: As GPUs become increasingly integral to high-performance computing and machine learning, ensuring memory safety in GPU programs has become crucial for reliable and secure execution. However, evaluating GPU memory safety techniques remains challenging due to the lack of comprehensive and standardized benchmarks. In this paper, we present GMSBench, a G
arXiv:2609.08968v1 Announce Type: new Abstract: Let $q=3^n$, where $n>1$ is odd, and let $g:\Fq\to\Fq$ be a perfect nonlinear (PN) function represented by a Dembowski--Ostrom (DO) polynomial. Put $\tau=g(1)$, let $\epsilon$ be the indicator of $\Fthree^*$, and, for $c\in\Fq$, define $\widetilde G_c(x):=g(x+c)+\tau\epsilon(x)$. We prove that every $\widetilde G_c$ is APN and has boomerang uniformit
arXiv:2609.08971v1 Announce Type: new Abstract: The rapid integration of AI into human-centred systems such as Brain-Computer Interfaces (BCIs) has created a poorly understood attack surface linking neural signals to physical systems. Exploits in this domain threaten cognitive autonomy, mental privacy, and physical safety, from neural data exfiltration to malicious control of BCI-tethered devices.
arXiv:2609.09087v1 Announce Type: new Abstract: As Large Language Model (LLM) agents increasingly automate offensive operations across the cyber kill chain, their efficacy in complex local post-exploitation tasks remains inadequately quantified. Among these, Linux privilege escalation is a key step between initial access and full system compromise. However, existing evaluations for this task are l
arXiv:2609.05430v1 Announce Type: cross Abstract: The effort to tokenize non-currency assets faces several hurdles, including the lack of a scalable decentralized computing infrastructure to manage asset-related metadata. While the centralized securities depository model has served the financial industry well for several decades, the vision of tokenization at a global scale requires new infrastruc
arXiv:2609.05503v1 Announce Type: cross Abstract: Diffusion models have become the dominant family of generative models in the visual domain. However, their widespread public availability enables misuse at scale, motivating a rapidly growing body of research on adversarial attacks and defenses. This survey provides, to our knowledge, the first unified review of this literature across three visual
arXiv:2609.05645v1 Announce Type: cross Abstract: We introduce ordered-angle coding for binary-unanimity testing among $n$ transformation-only users in a serial quantum architecture inherited from Loop-Back communication. User $B_i$ fixes one private sign across a logical word and applies $R(s_i\alpha_j)$ in trial $j$. If $w$ users choose the negative sign, serial composition gives $R[(n-2w)\alpha
arXiv:2609.05850v1 Announce Type: cross Abstract: Despite the significant efforts devoted to aligning large language models (LLMs) with human values and ensuring safe deployment, recent work has revealed that LLMs remain vulnerable to adversarial jailbreak attacks that can bypass safety guardrails and elicit harmful responses. Many defense methods are proposed to detect jailbreaks but they are lim
arXiv:2609.05889v1 Announce Type: cross Abstract: Resource-exhaustion attacks against autoregressive vision-language models (VLMs) typically assume unimodal threat models, treating the image branch as the primary optimization surface while holding user-visible prompts fixed. Even recent loop-centric variants remain confined to this single-channel paradigm, leaving the exploitation of availability
arXiv:2609.06346v1 Announce Type: cross Abstract: Continual learning (CL) enables models to acquire new knowledge from sequentially arriving tasks while retaining previously learned knowledge. However, in practical scenarios, task streams collected from untrusted sources may contain backdoor-poisoned samples, posing a critical challenge to the stability, plasticity, and security of continual learn
arXiv:2609.06415v1 Announce Type: cross Abstract: Encrypted quantum cloning enables the creation of multiple encrypted clones of an unknown quantum state while allowing only one effective decryption, with the decryption resource being intrinsically consumed in the process. In this paper, we argue that this property supports a distinctive custodial security primitive for post-compromise response. W
arXiv:2609.06500v1 Announce Type: cross Abstract: Agentic systems require secure and efficient delegation of privileges across multiple collaborating agents. Existing approaches fall into two categories. Some propagate user identities directly to agents, obscuring accountability and creating persistent over-privilege risks that are amplified by the non-deterministic behaviour of AI agents. Others
arXiv:2609.06872v1 Announce Type: cross Abstract: When a user asks an assistant to forget a record, the test is whether the memory now matches the state it would hold if the record had never been stored. Independently encoded rows can be removed directly; a recurrent memory folds records into an evolving state. One hope is a receipt: save the difference the record made when it arrived, carry it fo
arXiv:2609.06966v1 Announce Type: cross Abstract: Model misalignment, prompt injection, or operator misuse could lead AI agents operating frontier-lab accounts to exfiltrate model weights, poison training data, or weaken release gates. Existing benchmarks do not test whether defenders can detect this activity among routine work under a limited review budget. We introduce MOLE, an open benchmark of
arXiv:2609.07051v1 Announce Type: cross Abstract: World models increasingly serve as the predictive core of model-based reinforcement learning agents, enabling them to simulate future dynamics and reason over imagined trajectories before acting. Their substantial training demands make pretrained world models attractive for distribution and reuse, exposing downstream systems to model supply chain t
arXiv:2609.07162v1 Announce Type: cross Abstract: Several properties safety monitors are asked to certify, among them cross-tenant noninterference, sandbagging and evaluation awareness, are 2-safety hyperproperties, witnessed only by two executions. The standard consequence is a binary impossibility: one trace cannot decide them. We replace the binary with a measurement. A tight bound puts the bal
arXiv:2609.07224v1 Announce Type: cross Abstract: Security patches (SPs) are the main mechanism for fixing software vulnerabilities, yet a single vulnerability is not always resolved by a single patch: fixes may be completed incrementally, propagated across maintained branches, or replicated across related repositories. When patch records are incomplete, downstream users may observe only part of t
arXiv:2609.07312v1 Announce Type: cross Abstract: This paper proposes a robust decentralized personalized federated learning method R-DPFL, that enables clients to reduce the impact of Byzantine attacks via robust neighborhood direction estimation and history-based update trend prediction, rather than purely aggregating client models as in the existing work. In R-DPFL, each client first computes t
arXiv:2609.07675v1 Announce Type: cross Abstract: Transaction-local controls answer whether one financial request may proceed, but market behavior can be distributed across messages, agents, assets, and time. We study this interpretation gap in a virtual exchange populated by ten role-conditioned language-model agents. The agents communicate, trade reference assets and futures, launch tokens, and
arXiv:2609.07894v1 Announce Type: cross Abstract: The main goal of steganography is to transmit hidden messages in legitimate-looking communication messages. Phase-domain information hiding, however, has not been fully explored for quantum systems. This work introduces a finite-dimensional Quantum Hilbert Transform (QHT) as a unitary phase operator based on the Quantum Fourier Transform. Using thi
arXiv:2609.08062v1 Announce Type: cross Abstract: Tool-using language agents can delegate and revoke permissions while acting through external services. We show that two authorization histories can have identical current permissions and identical all-pairs reachability yet require opposite decisions after the same direct-edge revocation. We formalize the information needed to preserve such distinc
arXiv:2609.08331v1 Announce Type: cross Abstract: Video Large Language Models (VideoLLMs) are increasingly deployed in safety-critical applications such as content moderation and video analytics. To process long videos efficiently, VideoLLMs rely on frame sampling, token compression, and modality fusion, which together form an observation pipeline that reduces the raw video to a compact internal r
arXiv:2609.08342v1 Announce Type: cross Abstract: World models take multimodal inputs like text, photos, and diagrams to generate dynamic scenes in accordance with the laws of physics, thus opening a compelling application: fusing multimodal legal evidence to re-create a crime scene and re-enact how an offence could have been committed. However, feeding the raw, unorganized evidence into a world m
arXiv:2609.08373v1 Announce Type: cross Abstract: Aligned language models fail under two independent pressures: the structural jailbreak class recently formalized as Involuntary In-Context Learning (IICL), which reframes a harmful request as the final missing cell of a data-labeling task completed by pattern rather than judged as content; and the erosion of safety alignment outside English. A natu
arXiv:2406.15070v3 Announce Type: replace Abstract: We present Tempora-Fusion, the first homomorphic TLP scheme with efficient public verification of both individual puzzle solutions and homomorphic linear combinations. Tempora-Fusion lets clients generate puzzles independently, later authorize a linear combination with its own release time, and enables any party to verify the released result with
arXiv:2410.08776v3 Announce Type: replace Abstract: Large Language Models (LLMs) face significant security risks despite their advanced capabilities. While techniques like Reinforcement Learning with Human Feedback (RLHF) improve ethical alignment, excessive exposure to security-related training data may cause LLMs to overtrust such information, creating new vulnerabilities. Investigating this iss
arXiv:2503.00140v3 Announce Type: replace Abstract: Machine learning systems deployed in distributed or federated environments are highly susceptible to adversarial manipulations, particularly availability attacks -- rendering the trained model unavailable. Prior research in distributed ML has demonstrated such adversarial effects through the injection of gradients or data poisoning. In this work,
arXiv:2504.02109v2 Announce Type: replace Abstract: Cybersecurity incidents such as data breaches have become increasingly common, affecting millions of users and organizations worldwide. The complexity of cybersecurity threats challenges the effectiveness of existing security communication strategies. Through a systematic review of over 3,400 papers, we identify specific user difficulties includi
arXiv:2504.21323v3 Announce Type: replace Abstract: Knowledge distillation is widely used to transfer behavior from a large teacher model to a smaller student. It is often assumed to be safe when the teacher is clean, because classic backdoor attacks rely on poisoned labels and triggers in supervised training, whereas distillation trains the student to match a teacher's outputs. We show that this
arXiv:2509.11120v5 Announce Type: replace Abstract: Autonomous vehicles, including self-driving cars, ground robots, and drones, rely on multi-modal sensor pipelines for safe operation, yet remain vulnerable to adversarial sensor attacks. A critical gap is the lack of a systematic end-to-end view of how sensor-induced errors traverse interconnected modules to affect the physical world. To address
arXiv:2510.11570v3 Announce Type: replace Abstract: Open-weight Large Reasoning Models (LRMs) are approaching the capabilities of their frontier counterparts but pose significant safety concerns, as they are difficult to patch or monitor post-release. To prevent misuse, reasoning-based safety guardrails, where models explicitly reason on safety justifications before answering, have become a promis
arXiv:2510.11851v3 Announce Type: replace Abstract: We reveal that Deep Research (DR) agents systematically expose safety risks: simply submitting harmful queries that a standalone LLM would reject outright can elicit detailed and dangerous reports from DR agents. Empirical analysis reveals that the advantages that make DR agents powerful unintentionally make them vulnerable: both the research rol
arXiv:2510.16835v2 Announce Type: replace Abstract: Reliable family labels are essential for Android malware analysis, yet most widely used benchmarks derive such labels from aggregated VirusTotal engine outputs. Because antivirus vendors differ in detection logic, naming conventions, and signature updates, these labels are often inconsistent across engines and unstable over time, which introduces
arXiv:2511.08367v2 Announce Type: replace Abstract: Large Vision-Language Models (VLMs) are susceptible to jailbreak attacks: researchers have developed various attack strategies that bypass the safety mechanisms of VLMs. Among these approaches, jailbreak methods based on the Out-of-Distribution (OOD) strategy have garnered widespread attention due to their simplicity and effectiveness. This paper
arXiv:2512.24682v4 Announce Type: replace Abstract: The complexity, interdependence, and rapid evolution of 3GPP specifications present fundamental challenges for ensuring the security of modern cellular networks. Manual reviews and existing automated approaches, which often depend on rule-based parsing or small sets of manually crafted security requirements, fail to capture deep semantic dependen
arXiv:2604.03813v3 Announce Type: replace Abstract: Adams Bridge, a hardware accelerator for ML-DSA and ML-KEM designed for the Caliptra root of trust, masks 1 of its Inverse Number Theoretic Transform (INTT) layers and relies on shuffling for the remainder, claiming per-butterfly Correlation Power Analysis (CPA) complexities of 2^46 (ML-DSA) and 2^96 (ML-KEM). We evaluate these claims across seve
arXiv:2604.15845v2 Announce Type: replace Abstract: Modern computing systems implicitly trust human input devices, allowing USB Human Interface Device (HID) emulators, such as the USB Rubber Ducky, to inject arbitrary keystrokes while bypassing conventional defenses. Speed- and regularity-based heuristics are easily evaded through slower or randomized timing. Keystroke dynamics offers a behavioral
arXiv:2604.18510v2 Announce Type: replace Abstract: Open-weight language models can be rendered unsafe through several parameter-level interventions, yet models with matched harmful compliance can exhibit fundamentally different failure modes. We compare harmful supervised fine-tuning (SFT), harmful reinforcement learning with verifiable rewards (RLVR), and refusal-feature abliteration in Qwen2.5-
arXiv:2605.00487v2 Announce Type: replace Abstract: We introduce a technology to formally verify that a software system satisfies a temporal specification of functional correctness, without revealing the system itself. Our method combines a deductive approach to model checking to obtain a formal certificate of correctness for the system, with zero-knowledge proofs to convince an external verifier
arXiv:2605.07515v2 Announce Type: replace Abstract: Cybersecurity policy assessment is an evidence-intensive organizational decision task. Reviewers must locate relevant policy statements, determine whether they sufficiently address security controls, identify missing elements, and decide where further policy action is required. Existing AI-based approaches support parts of this process, but often
arXiv:2605.21694v2 Announce Type: replace Abstract: Connecting large language models (LLMs) to defensive enforcement requires more than asking a model whether an attack is happening. A defender must decide which model outputs may change the system state, which outputs must be rejected, and how failures should be recorded. We present PocketAgents, a manifest-driven library of autonomous defense age
arXiv:2605.29654v2 Announce Type: replace Abstract: The Manufacturer Usage Description (MUD) enables enforcement of network restrictions for IoT devices based on their expected network traffic, as specified by manufacturers in a MUD file. Devices advertise a URL pointing to this file, yet the standard does not define how to securely bind the issuing device to its profile. As a result, malicious de
arXiv:2606.05743v2 Announce Type: replace Abstract: Despite advances in safety alignment, large language models remain vulnerable to continuously evolving jailbreaks. Existing fine-tuned safety classifiers cannot adapt to these evolving attacks, while adaptive memory-based guardrails tend to over-refuse benign queries that resemble stored attacks. We propose Membrane, a self-evolving guardrail bui
arXiv:2606.30542v2 Announce Type: replace Abstract: The absence of authenticated bootstrapping between User Equipments (UEs) and Base Stations (BSs) in 5G leaves System Information Block (SIB) broadcasts unprotected, enabling fake BS attacks, man-in-the-middle interception, and spoofed emergency alerts. Prior efforts such as Public Key Infrastructure (PKI)-based certificate chains, token-based sch
arXiv:2608.00583v2 Announce Type: replace Abstract: Chain-of-thought (CoT) monitoring is meant to catch the reward hacks that look clean in the actions and betray themselves only in the reasoning. We show that this is exactly where an adversary who controls the reasoning can defeat it. Rewriting only an agent's reasoning to read as good-faith engineering, while copying every command and output ver
arXiv:2608.15475v2 Announce Type: replace Abstract: Quantized Vision-Language-Action (VLA) models expose a weight-fault surface: Rowhammer-style faults can corrupt deployed INT8 bits. We present the first bit-flip attack on a VLA: a few gradient-selected flips reduce closed-loop success to $0\%$, while hundreds of random flips are harmless. Across four model variants spanning three action-head fam
arXiv:2608.16403v2 Announce Type: replace Abstract: Process variables (PVs) provide the process evidence needed for process-aware security monitoring in industrial cyber-physical systems (CPSs). However, existing supervisory infrastructures expose only the subset of PV values recorded by historians, leaving many additional runtime PV values unobserved. To address this incomplete process visibility
arXiv:2504.08192v2 Announce Type: replace-cross Abstract: Machine unlearning is a promising approach to improve LLM safety by removing unwanted knowledge from the model. However, prevailing gradient-based unlearning methods suffer from issues such as high computational costs, hyperparameter instability, poor sequential unlearning capability, vulnerability to relearning attacks, low data efficiency
arXiv:2504.14957v3 Announce Type: replace-cross Abstract: Ma and Huang recently proved that the PFC construction, introduced by Metger, Poremba, Sinha and Yuen [MPSY24], gives an adaptive-secure pseudorandom unitary family PRU. Their proof developed a new path recording technique [MH25]. In this work, we show that a linear number of sequential repetitions of the parallel Kac's Walk, introduced by
arXiv:2512.13628v4 Announce Type: replace-cross Abstract: We study non-interactive zero-knowledge proofs (NIZKs) for NP satisfying: 1) statistical soundness, 2) computational zero-knowledge (ZK) and 3) certified-everlasting zero-knowledge (CE-ZK). The CE-ZK property allows a verifier of a quantum proof to revoke the proof in a way that can be checked (certified) by the prover. Conditioned on succe
arXiv:2606.14787v2 Announce Type: replace-cross Abstract: We study six production image-to-image AI systems (gpt-image-1, Gemini 2.5 Flash Image, Flux Kontext, SDXL img2img, SD3 img2img, and Qwen Image Edit) under a content-adaptive sub-JND adversarial perturbation pipeline, scoring all outputs by frozen DINOv2 ViT-B/14 token distances against clean references. Across a 3,588-call corpus spanning
arXiv:2608.12675v2 Announce Type: replace-cross Abstract: Retrieval-Augmented Generation (RAG) is widely used to improve the performance of Large Language Models (LLMs) in answering user queries. Existing privacy research on RAG has focused on preventing unauthorized users from accessing sensitive data. However, another important problem that is often overlooked in RAG privacy research is that ext
arXiv:2608.13010v2 Announce Type: replace-cross Abstract: Retrieval-augmented generation uses an external corpus as inference-time evidence, allowing an attacker to promote a false answer by injecting a handful of documents. Detection must distinguish this manipulation from ordinary relevance without knowing which queries or documents are targeted. Existing detectors use text irregularity, candida
arXiv:2609.05804v1 Announce Type: new Abstract: Why China Succeeds offers what no existing book does: a single, unified explanatory framework that traces China's rise from its ancient civilizational foundations to its twenty first century technological ascent. Rather than treating China's success as ideologically exceptional or historically accidental, it reveals the universal institutional logic
arXiv:2609.06085v1 Announce Type: new Abstract: Understanding the joint dynamics of prices and trades is central to market microstructure, where returns and order flow interact through nonlinear and state-dependent mechanisms. Linear models are interpretable but may miss these effects, while deep neural networks improve forecasting at the cost of transparency. We use neural networks as tools for s
arXiv:2609.06123v1 Announce Type: new Abstract: The relationship between investment policy associated with species growth profile is essential in seeking the most appropriate strategy for a policymaker. Balancing maximum profit with the sustainability of species remains a central issue in both ecological and economic contexts. This study presents a comparative analysis of two interest principles,
arXiv:2609.06137v1 Announce Type: new Abstract: Pathwise differentiation of Monte Carlo estimators fails at payoff discontinuities, producing zero or biased sensitivities for barriers, autocallables, and digital options. The industry workaround --- smoothing the indicator functions --- introduces bias and requires per-product calibration. We derive a correction formula that restores unbiased Greek
arXiv:2609.06422v1 Announce Type: new Abstract: We introduce ALM-GARCH, an asymmetric long-memory GARCH model in which positive and negative innovations enter conditional variance with different injection amplitudes and kernel offsets. These departures define testable level and memory channels relative to a nested symmetric benchmark. Positive Harris recurrence holds for interior configurations un
arXiv:2609.07011v1 Announce Type: new Abstract: We review the evidence on subsidized job training programs in industrial countries. We provide critical, narrative summaries of randomized studies of major job training in the US; of two "judge randomization" studies in the US and Denmark; and of previous meta-analyses. We then perform a meta-analysis of 56 randomized trials of job training in the US
arXiv:2609.07169v1 Announce Type: new Abstract: We develop semi-closed pricing formulas and lifted-model hedging methods for discretely monitored geometric and arithmetic Asian options in the Volterra-Heston stochastic volatility model. Exploiting the affine Volterra structure, we derive a tractable transform for the joint law of the terminal log-price and the discretely monitored geometric averag
arXiv:2609.07946v1 Announce Type: new Abstract: For more than four decades, the 60/40 stock/bond portfolio has served as a benchmark for delivering reasonable returns without excessive risk. More recently, a 50/30/20 stock/bond/alternative portfolio has been suggested. We use gold as the alternative and as an inflation hedge. In this paper we ask: how much improvement over these benchmark fixed-we
arXiv:2609.07958v1 Announce Type: new Abstract: This text grew out of a historical introduction initially written for a study of interest rates in cryptocurrency markets. The difficulty of defining a term structure for a currency without a conventional bond market led naturally to a more fundamental question: under what historical conditions does a yield curve become observable at all? Credit exis
arXiv:2609.08198v1 Announce Type: new Abstract: Bouchaud and Farmer (2023) argue that self-fulfilling beliefs generate a realistic wealth distribution. Their reported Gini coefficient, quantiles, and Pareto exponent are computed from dynasty-level wealth averaged over 250 dates, not from a cross section. Correcting the estimand strengthens their qualitative result: with $N=1{,}000{,}000$ dynasties
arXiv:2609.08541v1 Announce Type: new Abstract: We study variance-optimal stock hedging and the convergence of approximate strategies in the rough Hawkes--Heston model. Starting from the model's affine conditional transform and the affine Volterra jump framework, we obtain semi-explicit hedges for European calls and a representation of the minimum quadratic error through the Galtchouk--Kunita--Wat
arXiv:2609.08605v1 Announce Type: new Abstract: Let \(P\) be a fixed physical law and let \(Q\) be an equivalent martingale measure selected from the martingale-measure set associated with a chosen numeraire. A change of numeraire maps \(Q\) to \(T_LQ\), where \(d(T_LQ)=L\,dQ\) and \(L\) is the terminal likelihood ratio. The forward relative-entropy projection minimizing \(D_{\mathrm{KL}}(P\Vert Q
arXiv:2609.08881v1 Announce Type: new Abstract: We develop a theoretical model that endogenizes the regulator's decision to impose short-selling bans to prevent large stock price declines. Empirically, we test the model's predictions using the cross-sectional variation in short-selling restrictions implemented across European countries in 2020. Consistent with our model, we find that bans had a de
arXiv:2609.03654v1 Announce Type: cross Abstract: The comparative analysis of banks' financial statements poses significant challenges for automated question answering systems due to their complexity, substantial length, technical language, and inhomogeneity of both textual and numerical content across different jurisdictions and institutions. We introduce FinRAG-QA, a novel benchmark dataset for
arXiv:2609.06739v1 Announce Type: cross Abstract: Sports betting moves money continuously from a large public to a small number of firms. The most influential account of that flow, due to Levitt (2004), holds that books price away from the market-clearing point to exploit predictable public biases. It computes profit from two numbers (the probability that a side wins the proposition and the fracti
arXiv:2609.07207v1 Announce Type: cross Abstract: We develop a filter for time series, defined at each time $t$ as the minimizer of a discounted convex combination of observed and expected losses. The filter can be estimated by simulation to an arbitrary level of accuracy in $O(1)$ flops at each time point $t$ and can be run for all values $t=1,...,T$ in parallel. These methods are applied to robu
arXiv:2609.07351v1 Announce Type: cross Abstract: We establish existence, uniqueness, stability, and convergence results for one-dimensional $q$-Bass martingales, characterized as the martingales with prescribed initial and terminal marginals whose transition kernels are closest to a reference measure $q$. Their existence is equivalent to the solvability of a fixed-point problem for probability di
arXiv:2609.07731v1 Announce Type: cross Abstract: We show that ordinary business language --- "maximize profitability" --- induces profit-oriented ambiguity resolution: LLMs systematically dismiss ambiguous signals of potential safety violations to serve business objectives. In 3,600 controlled trials across eight reasoning-capable LLMs, adding a profit mandate to otherwise identical prompts incre
arXiv:2609.08060v1 Announce Type: cross Abstract: We build and evaluate a pre-game win-probability forecaster for individual maps (``games'') in professional \emph{League of Legends} (LoL). The proposed model is a one-stage logistic regression fit end-to-end on the win/loss log-loss: each team's exponentially-weighted moving average of past same-side results, a ridge-shrunk stable strength that is
arXiv:2609.08106v1 Announce Type: cross Abstract: MASTER's inter-stock multi-head attention -- the module responsible for modeling cross-sectional stock relationships -- accounts for 42.5% of model parameters and 25% of predictive value. We systematically decompose this module and uncover a surprising structure: the learned attention is near-uniform (perplexity 278/300), yet forcing exact uniformi
arXiv:2609.08581v1 Announce Type: cross Abstract: Formulaic alpha discovery is a pool-dependent symbolic search problem in which informative feedback is observed primarily when a complete expression is evaluated. This delayed feedback creates two coupled difficulties: the retained alpha pool does not preserve the full history of realized evaluation feedback, and the value of an intermediate constr
arXiv:2609.08929v1 Announce Type: cross Abstract: As US power markets contend with growing demand for firm generation, the nuclear industry has offered Small Modular Reactors (SMRs). However, how these concepts would fare in a rapidly evolving power grid is unclear, given the paucity of operational examples. Current literature, informed by substantial cost escalations for traditional nuclear plant
arXiv:2506.08067v4 Announce Type: replace Abstract: We investigate the asymptotic behaviour of the Bachelier implied volatility tails, extending the large-strike results established for the Black-Scholes implied volatility. Exploiting the theory of regular variation, we derive explicit expressions for the Bachelier implied volatility in the wings of the smile, directly linking them to the tail dec
arXiv:2511.04782v2 Announce Type: replace Abstract: Depending on the persistence of a one-off shock bringing the economy to the Effective Lower Bound (ELB), the standard New Keynesian model predicts starkly different conclusions. We offer a potential solution to this morass by assuming that the one-off shock is such that the economy necessarily leaves the ELB in finite time. Under these assumption
arXiv:2511.15364v2 Announce Type: replace Abstract: Anonymizing financial texts prevents large language models (LLMs) from exploiting look-ahead bias, but inadvertently weakens the extracted signal. We propose a framework disentangling this information loss from bias removal. Predicting S&P credit downgrades, raw texts significantly outperform anonymized texts. Look-ahead bias does not explain thi
arXiv:2512.07840v2 Announce Type: replace Abstract: Since its inception, Bitcoin has been positioned as a revolutionary alternative to national currencies. This paper evaluates that claim against two competing theoretical traditions -- Post-Keynesian monetary theory and the Austrian School -- while evaluating the strongest counter-arguments within each tradition rather than treating either verdict
arXiv:2512.23515v2 Announce Type: replace Abstract: Signal decay and regime shifts pose recurring challenges for data-driven investment strategies in non-stationary markets, where conventional time-series and machine learning approaches often struggle to generalize beyond historical correlations. While large language models (LLMs) offer strong capabilities for processing unstructured information,
arXiv:2603.24605v2 Announce Type: replace Abstract: Martingale Optimal Transport (MOT) provides a framework for robust pricing and hedging of illiquid derivatives. Classical MOT enforces exact calibration of model marginals to the mid-prices of vanilla options. Motivated by the industry practice of fitting bid and ask marginals to vanilla prices, we introduce a relaxation of MOT in which model-imp
arXiv:2604.09871v3 Announce Type: replace Abstract: I develop a model in which occupational knowledge links the organization of production to democratic performance. In production, specialized knowledge allows workers to perform narrow tasks efficiently, while cross-field knowledge allows coordinators to keep the system coherent. In politics, voters bring their occupational knowledge to the ballot
arXiv:2604.18118v4 Announce Type: replace Abstract: Can contagion components be identified in aggregated default counts when default probabilities fluctuate with the economic environment? We study this question as an identifiability problem for coarse-grained default-count distributions. We adopt a binomial mixture with a Probit--Normal default probability as a parsimonious null model for environm
arXiv:2605.06604v2 Announce Type: replace Abstract: This paper proposes a hybrid methodology to improve the approximation of SABR (Stochastic Alpha Beta Rho) implied volatility by combining analytical structure with machine learning. The approach augments the neural-network input representation with geometric features derived from the stochastic differential equations of the SABR model. Unlike app
arXiv:2606.14182v3 Announce Type: replace Abstract: We give a unified analytic account of correlation emergence and the Epps effect in two coupled limit order books. The Epps effect is the empirical reduction in measured cross-asset log-return correlation at short aggregation scales, or equivalently the recovery of measured correlation as the aggregation interval increases. The model starts from a
arXiv:1510.03928v5 Announce Type: replace-cross Abstract: This work is motivated by numerical solutions to Hamilton-Jacobi-Bellman quasi-variational inequalities (HJBQVIs) associated with combined stochastic and impulse control problems. In particular, we consider (i) direct control, (ii) penalized, and (iii) semi-Lagrangian discretization schemes applied to the HJBQVI problem. Scheme (i) takes th
arXiv:2608.00045v2 Announce Type: replace-cross Abstract: This study shows that textual descriptors alone can predict early-stage startup success, defined as Exit, without relying on contextual, financial, or human capital variables. Using venture capital-curated datasets covering 7,419 startups over 20 years, the research isolates text-based framing variables and engineers 850 features through st
arXiv:2608.13732v2 Announce Type: replace-cross Abstract: Using the local time-space calculus of Peskir (2005) and the method developed in Mijatovic (2010), we derive a new integral representation for the distribution of the first-passage time (FPT) of a diffusion process through a time-dependent barrier. We present a complete three-step numerical algorithm: first, the problem is reduced to a Volt
OT/IT in a nutshell. This humorous illustration, created in the style of an *Asterix* comic, parodies IT and OT environments by depicting technological concepts as a Gallic village that must defend itself against external threats like ransomware. The composition is designed as a bird's-eye view packed with detail, revealing a world divided in two: a sprawling IT compound occupies the upper section
Colonel Anurag Upadhyay, a former Indian Army Special Forces officer, is seeking expert medical and technological support after suffering a severe stroke in August 2025 that resulted in locked-in syndrome and bilateral below-knee amputations. Despite remaining conscious and communicating through eye movements, his family is exploring advanced rehabilitation, neurological research and assistive tec
# TIL (Today I Learned) why the sampling rate of CDs is 44,100 Hz. It has a lot to do with video resolutions. In the 1970s audio engineers wanted to create digital audio masters, since analog masters had noise problems. But no digital medium large enough to store music existed. So they exploited the television broadcast standards to store their audio samples on the available tapes albeit designed
There are troubling new signs that the overhaul of Medicare drug coverage in the Inflation Reduction Act is dramatically driving up program spending. Why it matters: The upward trajectory could threaten some seniors' coverage for outpatient prescription drugs — and it could force painful tradeoffs over the next decade as Medicare consumes a growing share of the nation's debt. Follow the money: The
Iran is rapidly building out a suspected nuclear facility buried deep inside a granite mountain near Natanz — dubbed "Pickaxe Mountain" — with satellite imagery showing a construction surge in 2026. The site is likely designed to shelter centrifuges or enrichment work beyond the reach of current U.S. bunker-busters. Trump has threatened to strike it ("we may hit Pickaxe very soon"), but the Pentag
SAN JUAN, Puerto Rico – On September 8, 2026, Alexander Ortiz-Robles, a Puerto Rico Department of Treasury (Departamento de Hacienda or “PRDT”) employee was charged via information and pleaded guilty to participating in a bribery conspiracy involving the loss of approximately $6,772,578 in Puerto Rico tax revenue, in violation of 18 U.S.C. § 371. Ortiz-Robles is scheduled to be sentenced before U.
Colin the crab lived a simple life. A happy life. He was engaged to Lily the lobster, and they were very much in love. However Luke, king of the lobsters, intervened. "I will not have my daughter marrying a creature that walks sideways." Luke said. "It would bring shame upon the whole lobster kingdom." Lily cried and begged her father to see reason, but to no avail. Distraught, Colin scuttled side
CVE-2026-67401 is a nasty one for cPanel/WHM operators. cPanel disclosed an SQL injection in its EmailTrack functionality that allows an authenticated account with mail-related privileges to create arbitrary files on the server. Successful exploitation can lead to code execution as root, giving an attacker full control of the server. All supported cPanel/WHM versions are affected. Patched builds:
🚨 𝗜𝗻𝘀𝗶𝗱𝗲 # 𝗧𝗲𝗿𝗺𝗶𝗻𝗮𝗹𝗙𝗶𝘅 : 𝗪𝗼𝗿𝗱-𝗘𝗻𝗰𝗼𝗱𝗲𝗱 𝗣𝗮𝘆𝗹𝗼𝗮𝗱𝘀, 𝗦𝗺𝗮𝗿𝘁-𝗖𝗼𝗻𝘁𝗿𝗮𝗰𝘁 𝗟𝘂𝗿𝗲𝘀, 𝗙𝗼𝗿𝘂𝗺-𝗕𝗮𝘀𝗲𝗱 𝗖𝟮 ⚠️ Targeting the US and Canada, the chain starts on a compromised WordPress site and ends with a WinHTTP stager running inside a signed Microsoft executable. A user-driven lure becomes trusted-process C2 activity that can evade early validation and
It's interesting to wonder if an island could be self-sufficient forever in the face of ever-increasing consumption. Also, bluetongue? Not familiar with it. Not good! Viral. I don't know. Despite my last name, I'm not a farmer. All I know is what I read in the papers. I would say, as a species, we've a long row to hoe on this blue sphere we call home. "The appeal coincides with a new NFU study rev
Manon Kerjean, FRANCE 24's film critic highlights some of her picks among the French films in competition at the 83rd Venice International Film Festival. She also looks at new releases in French cinema: "15/18", a film set in a psychiatric unit which tackles mental health and "A Woman's Life" – a film starring Cesar-winning actress Léa Drucker as a woman whose organized life is upended by the arri
Context-Masked Truncated Reasoning Audits for Answer-Key Dependence in LLM Tutors Large language model (LLM) tutors may have access to teacher notes, answer keys, rubrics, or retrieved solutions while producing student-facing explanations. We study whether truncated reasoning probes can distinguish direct access to such private context from… arXiv — Computers & Society + AI · cs.CY, cs.AI updates
Heiko Tietze has published a blog post summarizing the results of a recent survey about the use of LibreOffice's database application, Base. 455 people participated in the survey, including more than 330 who use Base on Linux, with use cases ranging from maintaining records of personal media such as CDs or DVDs to use enterprise-resource planning (ERP) and finance. Of course, users had many ideas
Save the Dates 📌 Die Sommerferien neigen sich dem Ende zu und wir legen wieder los! Die nächsten drei OWASP Heilbronn Stammtische stehen fest: 📅 Donnerstag, 08.10.2026, 18 Uhr 🎤 "An SSDLC Journey: A Story with Rats and Sea Life" – Henrik Willert 📅 Donnerstag, 05.11.2026, 18 Uhr 🎤 "Pentesting mit KI – Zwischen Schlagzeilen und Realität" – Christian Biehler 📅 Donnerstag, 10.12.2026, 18 Uhr 🎤
Hi HN, I am Dimitris. This is a company OS that I built and use to run my business and anyone can install it and self host it for free. Think of it as Claude Code, Cowork and the cloud sessions in one self hosted application. It is a Multi-tenant application by design where many people can collaborate on the company agents with 4 different modes of collaboration, and it runs with your Anthropic or
The counterattck north of Lyman is a small area in the grand scheme. And if you're looking for signficance, you shouldn't over react to the push-pull of front lines in small sectors. But - there is maybe a story behind the story here. Several stories. - The closest you're going to get to The Truth and an honest insight of Russian thinking and awareness is probably an aggregate of similarly timed a
Ivanti Patches Eight Critical and High-Severity Flaws in Neurons for ITSM Ivanti patched eight critical and high-severity vulnerabilities in Neurons for ITSM, including unauthenticated remote code execution flaws discovered via AI-driven security scanning. **If you run Ivanti Neurons for ITSM on-premises (versions 2025.2 through 2026.1), download and apply the September 2026 security patches from
"As horrific as it is to watch the World Trade Center towers crumble to the ground, those indelible images were recorded from the relative safety of the streets below. The final phone calls and voicemails left by 9/11 victims, however, provide a raw, real-time record of what it was like to be trapped inside the towers or aboard one of the hijacked planes. "'We so often look back at 9/11 knowing ho
‼️ LAPSUS$ Group has a new PGP signed message. Chapter II first victim is set to release in 3 days. IOC: lapsus[.]ar[.]io "When we completed our first operations, we demonstrated the triviality of modern enterprise defense systems. To TeamPCP, we send our sincere gratitude: your sacrifice provided us with the exact coverage we needed. We owe you all our gratitude for falling and taking on the acts
🚨🇹🇭 Alizune database and source code allegedly leaked on a cybercrime forum ⠀ Alizune is a Thailand-based online marketplace selling digital accounts and other products. ⠀ Actors identifying themselves as ZentinelTeam, including Keishell and noname8173, claim to have obtained and leaked the platform’s database and website source code. ⠀ Claimed exposed data includes: ⠀ • Usernames and email add
arXiv:2609.09212v1 Announce Type: new Abstract: This paper presents an end-to-end evaluation framework for image-triggered command injection against computer-use agents (CUAs). The goal is to test whether a local visual patch can induce verifiable environmental consequences along the full chain of screenshot input, VLM generation, action parsing, and environment execution. We train and deploy patc
arXiv:2609.09345v1 Announce Type: new Abstract: Off-path network attacks against TCP and DNS (over UDP) client-server connections are generally considered impractical nowadays, due to built-in security features in these protocols, e.g. randomized TCP (initial) sequence numbers and randomized UDP source ports, respectively. In this work, we refute this presumption by demonstrating that an unprivile
arXiv:2609.09357v1 Announce Type: new Abstract: Fully homomorphic encryption (FHE) enables inference on private data without revealing it to the server, but evaluating an entire input under FHE is expensive. We study \emph{selective homomorphic inference}, where only a sensitive region of interest (ROI) is encrypted, and computations independent of that region are performed in plaintext. Selective
arXiv:2609.09404v1 Announce Type: new Abstract: Agentic AI frameworks let a language model plan, keep memory, and call tools that reach real files, mail, and services. Most of these agents also read images, which gives an attacker a way to put text into the agent's context without going through the user. We present MMPIBench, a reproducible benchmark that measures what happens next. It delivers a
arXiv:2609.09420v1 Announce Type: new Abstract: Full-duplex speech models accept user speech while generating responses, creating an underexplored attack surface. We introduce DuplexJail, which delivers fixed, request-independent spoken prompts through the user audio channel. We compare fixed-delay interruption after the harmful request ends with refusal-triggered interruption following a cue in t
arXiv:2609.09423v1 Announce Type: new Abstract: Fully Homomorphic Encryption (FHE) enables privacy-preserving cloud services but incurs substantial computation overhead, making hardware acceleration essential. Among FHE operations, key-switching is a major performance bottleneck. Recent cryptographic advances introduce a novel key-switching method (i.e., KLSS) that reduces certain operational comp
arXiv:2609.09551v1 Announce Type: new Abstract: Recommender systems have become core infrastructure for modern online platforms, personalizing content at scale and strongly influencing what users see, click on, and purchase. However, this dependence on user interaction also exposes them to shilling attacks, where malicious actors can inject fake profiles to distort item rankings and control visibi
arXiv:2609.09553v1 Announce Type: new Abstract: Large language model safety and security research is preoccupied with, among other things, detecting and preventing jailbreak attacks: alignment bypasses that allow an adversarial user to elicit unwanted or harmful outputs from models. Arbitrary cipher, or covert communication, attacks are one such type of jailbreak and have previously been demonstra
arXiv:2609.09623v1 Announce Type: new Abstract: Byzantine agreement (BA) is a foundational building block in distributed systems, and the security analysis of BA protocols under multi-instance executions has attracted increasing attention. However, most existing adversary models focus solely on party corruption and neglect important threats posed by adversarial manipulations of communication chann
arXiv:2609.09697v1 Announce Type: new Abstract: Five years after the enforcement of the California Consumer Privacy Act (CCPA), understanding how website privacy practices evolve at scale in response to regulation remains a key challenge for both researchers and regulators. Prior work and regulatory efforts have focused on manual and case-specific enforcement, but there remain no scalable approach
arXiv:2609.09793v1 Announce Type: new Abstract: Directional ablation removes an aligned language model's ability to refuse by projecting a single "refusal direction" out of the weights that write the residual stream. It needs no gradient-based training and no optimization, only a few hundred contrastive prompts, which makes it the canonical white-box attack on open-weight alignment. However, it ha
arXiv:2609.09798v1 Announce Type: new Abstract: Large language models (LLMs) have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu- rity remains unclear. We introduce CS-Guard, the first benchmark to systematically evalu- ate guardrails for code generation security. It covers 1) text-to-code generation with 1000 high-quality malware-generation pr
arXiv:2609.09807v1 Announce Type: new Abstract: The integration of the Industrial Internet of Things (IIoT) into manufacturing has transformed industrial operations by optimising production management and ensuring product quality through smart industrial sensors that regulate processes based on real-time data. However, these sensor nodes are highly vulnerable to cyber threats, posing significant s
arXiv:2609.09817v1 Announce Type: new Abstract: Zonal in-vehicle networks ship Ethernet, MACsec, and TSN, but treat the network itself as trusted: once configured at the factory, there is no standardized runtime way to easily revoke access, rotate keys, or contain a compromised ECU. Zero Trust Architecture targets exactly that gap, yet existing automotive ZTA proposals bolt on dedicated infrastruc
arXiv:2609.09848v1 Announce Type: new Abstract: Synthetic data releases are increasingly proposed in the literature as a means of sharing realistic data replicas in lieu of sensitive private datasets. Even when the worst-case privacy leakage of such releases is bounded by means of differential privacy (DP), in practice a residual risk remains. Membership inference attack (MIA) audits are conducted
arXiv:2609.09963v1 Announce Type: new Abstract: Smartphones simultaneously transmit temporary identifiers over LTE, WiFi, and BLE. Existing privacy defenses analyze identifier randomization per protocol, implicitly assuming that these protections compose across protocols. We show that they do not: Even when each protocol leaks only temporary identifiers and the adversary is fully passive, unsynchr
arXiv:2609.10117v1 Announce Type: new Abstract: Trusted Execution Environments (TEEs) offer a promising mechanism for safeguarding the intellectual property of on-device Large Language Models (LLMs). To overcome the inherent computational bottlenecks of TEEs, existing TEE-Shielded LLM Partition (TSLP) methods apply efficient obfuscation schemes to computationally intensive layers, offloading them
arXiv:2609.10246v1 Announce Type: new Abstract: Once complex, creating and deploying a new cryptocurrency has become trivial. Coin launchpads now allow users to generate a new coin with merely a few clicks, at a minimal cost. Launchpad popularity has grown in tandem with the rise of "meme coins," which usually do not offer any novel technological properties and are purely created for fun. The most
arXiv:2609.10264v1 Announce Type: new Abstract: Open-source large language models (LLMs) are increasingly competitive with closed-source models while offering transparency and the ability to run inference without exposing user inputs to a service provider. However, running large-scale models locally requires substantial computational resources. In practice, users may still resort to a third-party
arXiv:2609.10294v1 Announce Type: new Abstract: ReDoS vulnerabilities are a type of denial of service software weakness that occurs when a regex is used to validate user-supplied input. In some cases, the regex matching process can take exponential time, leading to a denial of service. In this study, we examine and compare the effectiveness of five publicly-available regex detection tools, and one
arXiv:2609.10298v1 Announce Type: new Abstract: Cyberattacks against Operational Technology (OT) systems, which monitor and control industrial processes, pose an increasing threat to essential societal services. For this reason, developing automated intrusion response strategies is highly important. In this paper, we present a formal model of an OT intrusion response use case using the POMDP frame
arXiv:2609.10416v1 Announce Type: new Abstract: Watermarking the final patch produced by a coding agent provides provenance evidence for the submitted artifact, but does not authenticate the visible process that produced it. Behavioral watermarking methods primarily provide a global detection or identifier-recovery signal, so a locally edited trajectory may retain sufficient ownership evidence wit
arXiv:2609.09161v1 Announce Type: cross Abstract: The deployment of AI-driven Digital Twins (DTs) in large-scale Internet-of-Things (IoT) ecosystems demands continuous, high-fidelity synchronization between the physical environment and its virtual replica. Conventional approaches rely on dense sensor deployments, which introduce prohibitive costs in terms of hardware, energy, and network bandwidth
arXiv:2609.09334v1 Announce Type: cross Abstract: A fault-tolerant quantum computer runs behind a telemetry stream logging syndromes, decoder actions, resets and timing separately from the answer. Can it reveal the logical input? For a distance-$d$ rotated surface-code memory on a fixed schedule of $T=\Theta(d)$ rounds, under three stated hypotheses (sector-scalar honest backbone, transcript local
arXiv:2609.09511v1 Announce Type: cross Abstract: Physical-layer authentication (PLA) in non-terrestrial networks (NTNs) is challenged by severe Doppler shifts, long delays, and fast channel variations, which cause distribution shifts and degrade conventional learning methods. Existing PLA schemes often rely on single features or generalize poorly to unseen environments. This paper proposes a secu
arXiv:2609.09536v1 Announce Type: cross Abstract: Average treatment effect (ATE) estimation in observational studies is a fundamental statistical tool used frequently in social science, medicine, and other fields. These fields often work with sensitive data where privacy protections are important, so a differentially private mechanism for ATE estimation is highly desirable. Here we present two pro
arXiv:2609.09564v1 Announce Type: cross Abstract: The digitalisation of electrical distribution networks has increased the exposure of power-grid infrastructure to cyber attacks. Existing intrusion detection systems (IDSs), however, often rely on computationally expensive deep learning models that are difficult to deploy at the edge. Periodic retraining also exposes these systems to machine unlear
arXiv:2609.09659v1 Announce Type: cross Abstract: Federated learning shares model updates rather than raw data, yet these updates can be inverted to reconstruct the clients' training data. Analytic reconstruction attacks, which invert a gradient in closed form, degrade as the batch grows: prior single-round attacks recover only about half of a batch of size $100$ even when the attacker fully contr
arXiv:2609.09838v1 Announce Type: cross Abstract: Proximity gaps are central to the soundness of interactive oracle proofs of proximity (IOPPs) and polynomial commitment schemes (PCSs). An $[n,k,d]$ linear code $C\subseteq\mathbb F^n$ has a $\delta$-proximity gap with error $\epsilon$ if, for every $u_0,u_1\in\mathbb F^n$, either all points on $\ell_{u_0,u_1}=\{u_0+\alpha u_1:\alpha\in\mathbb F\}$
arXiv:2609.10002v1 Announce Type: cross Abstract: Deepfake detectors remain vulnerable to transfer-based black-box attacks, in which adversarial examples are generated on a source surrogate model and transferred to a target model, unknown to the attacker. Yet how source--target compatibility shapes attack success remains poorly understood. Prior studies evaluate limited detector pools and rarely d
arXiv:2609.10347v1 Announce Type: cross Abstract: Data-intensive applications move large amounts of data from storage to the compute unit, incurring significant data movement overhead. Storage-centric computing reduces this overhead by moving computation near or inside solid-state drives (SSDs). Enabling it requires modifying SSD policies, e.g., address translation and garbage collection, which ar
arXiv:2609.10412v1 Announce Type: cross Abstract: Static analysis remains a cornerstone of software security, yet the effectiveness of tools such as CodeQL is often limited by the substantial manual effort required to develop high-coverage query suites. While large language models (LLMs) have emerged as a potential solution for automated code reasoning, their practical utility in generating struct
arXiv:2509.22857v3 Announce Type: replace Abstract: Recent work using Fully Homomorphic Encryption (FHE) has made non-interactive privacy-preserving inference of deep Convolutional Neural Networks (CNNs) possible. However, the performance of these methods remains limited by their heavy reliance on bootstrapping, a costly FHE operation applied across multiple layers that severely slows inference. M
arXiv:2607.03215v2 Announce Type: replace Abstract: Generative models now write application code, harden and monitor it, and probe it for exploitable flaws, so that one family of models increasingly plays builder, defender and breaker at once. The prevailing view treats full autonomy as the natural end point of assistance. This article argues for a narrower and more defensible position than a blan
arXiv:2608.00375v2 Announce Type: replace Abstract: Recent work on Grafting decouples scale factors from ciphertext moduli, enabling more flexible precision management in RNS-CKKS. However, the multiplicative depth remains fundamentally constrained by the modulus chain structure. In this paper, we propose \emph{Trimming}, a novel fine-grained level management mechanism that decouples multiplicativ
arXiv:2608.01454v3 Announce Type: replace Abstract: Provenance-based intrusion detection systems (PIDS) frequently report strong performance, but the conclusions drawn from these results can be highly sensitive to benchmarking choices and evaluation protocols. We investigate this dependency by re-evaluating representative PIDS on public datasets that meet our audit, labeling, and calibration requi
arXiv:2511.13893v2 Announce Type: replace-cross Abstract: In differentially private (DP) tabular data synthesis, the consensus is that statistical models are better than neural network (NN)-based methods. However, we argue that this conclusion is incomplete and overlooks the challenge of densely correlated datasets, where intricate dependencies can overwhelm statistical models. In such complex sce
arXiv:2602.18934v3 Announce Type: replace-cross Abstract: Membership inference attacks (MIAs) threaten the privacy of machine learning models by revealing whether a data point was used during training. Existing MIAs often assume access to public datasets, shadow models, confidence scores or the training distribution, which makes them vulnerable to defenses like confidence masking. Label-only MIAs
arXiv:2609.09337v1 Announce Type: new Abstract: Whatever the policy question under consideration, reasoned and realistic evaluation requires credible policy analysis under uncertainty. This holds particularly to the study of climate policy in the United States, where science and ideology have become increasingly entangled. Welfare economics provides a transparent formal framework to describe and e
arXiv:2609.09470v1 Announce Type: new Abstract: We study how to combine diffusion models that disagree about drift and covariance. Candidate-first relative-entropy minimization gives geometric pooling, whereas expert-first minimization gives the arithmetic mixture associated with weighted logarithmic wealth. Different quadratic variations can make path-space entropy infinite, and the arithmetic mi
arXiv:2609.09673v1 Announce Type: new Abstract: Drug-drug interaction (DDI) errors pose serious risks to patient safety. Existing decision-support systems often require physicians to respond to alerts, disrupting workflows and contributing to high override rates. We examine whether a non-mandatory information intervention can reduce DDI errors and foster learning. Using a randomized field experime
arXiv:2609.10053v1 Announce Type: new Abstract: In many power systems, wind and solar generation increasingly often exceeds electricity demand. Curtailing renewable generation in those hours matters both for prices and for the physical stability of the grid. Turning off wind turbines and solar panels is technically easier than ramping down a large power station, yet support schemes often give rene
arXiv:2609.10380v1 Announce Type: new Abstract: Whether organisational separation of elective and emergency care improves physician productivity remains an open question. Most existing evidence relies on cross-sectional comparisons or volume-based outcomes that cannot isolate efficiency gains from input expansion or provider selection. In contrast, this paper exploits the staggered rollout of NHS
arXiv:2609.09588v1 Announce Type: cross Abstract: Signal correlation and PnL correlation are correlations over different index sets - across assets at each date versus across dates for scalar payoffs - and practitioners often treat the first as a proxy for the second. We give an exact decomposition that shows what that proxy sees and what it discards. We recall that at each date a normalized signa
arXiv:2609.10074v1 Announce Type: cross Abstract: We investigate the strategic communication of narratives under model uncertainty. The sender has private information about the true data-generating process of publicly observable data. The receiver is uncertain about how to interpret the data, but aware of the sender's incentives to strategically provide interpretations (``narratives''). We theoret
arXiv:2609.10227v1 Announce Type: cross Abstract: We present a comprehensive description of the network of Italian Business-to-Business commercial relationships, based on the universe of electronic invoices collected by the Italian Tax Office. The firm-to-firm detail of the data is exploited to describe the distribution of the numbers of buyers (customers) and sellers (suppliers) per firm, the cen
arXiv:2502.21141v3 Announce Type: replace Abstract: We examine how railroads shaped Denmark's nineteenth-century economic transformation and the diffusion of Grundtvigian institutions. Using a panel of 1589 parishes and a difference-in-differences design for staggered adoption, we find that railroad connection increased local population by 7 percent, partly through internal in-migration, and struc
arXiv:2509.24747v2 Announce Type: replace Abstract: Risk measures summarize the risk profile of financial positions in a single metric, which allows their comparison and supports investment decisions. When they respect common stochastic orders such as second-order stochastic dominance (SSD), they rest on sound conceptual grounds. Since SSD can be overly conservative and rigid, we study Meyer risk
arXiv:2607.05320v3 Announce Type: replace Abstract: Conditional portfolio models estimate risk relative to a chosen information set, yet rarely test whether that information removes common cross-asset dependence. When it does not, systematic risk may be treated as idiosyncratic, distorting portfolios and attainable efficient frontiers. We formulate this prior problem as screening-off for portfolio
arXiv:2601.09673v3 Announce Type: replace-cross Abstract: All existing match classification models in the tournament design literature suffer from two major limitations: a contestant is considered indifferent only if uncertain future results do never affect its prize, and competitive matches are not distinguished with respect to the incentives of the contestants. We propose a probabilistic framewo
Russia and Vietnam signed a package of energy-related cooperation documents during Vietnamese leader To Lam’s state visit to Moscow, where he held talks with Russian President Vladimir Putin at the Kremlin on September 9 [gov.ru#1][vietnamplus.vn#1][vietnamplus.vn#2][nhandan.vn#2]. Russia’s Energy Ministry said the documents included a framework agreement for cooperation on a strategic oil reserve
I wanted another computer as a sandbox, I don't really trust VMS, and fucking chip shortages, so to the boneyard it is. I have a disused ASRock Rack J1900D2Y with 8GB of RAM and a PicoPSU that used to be my mail server. Somehow, at some point, I modded the 1U case to mount the rack ears on the back. Annoying. So I dig up the original rack ears which I apparently still have, and try to flip things
I’m seeing some developers negatively freaking out over the iPhone Duo, because it’s a whole new design with very little lead time. Not just app developers but website developers too. I’m going to suggest that maybe you can just… Not. You are allowed to not let a massive greedy corporation impose themselves on your sanity because they wanted to make more money with yet another unnecessary, derivat
Give an AI agent a mere rumor of an exploit, and it’s enough for them to find it. What’s worse, I found I could use my own agents to find the exploit just by knowing roughly what it was about and so could have been exploiting it well before the public patch was available! Given that just the rumour of a security issue seems enough to give attackers enough info to find new exploits, we’re going to
The car in front if me struck a young cat and killed it. I only stopped because I saw it move... weakly. Poor thing was a teenager, gray, white. The car did not stop. It was still warm, it evacuated its bladder, but other than that, it was integral. I could feel no hearbeat and no pupil response. I took it to the emergency vet, because I was hoping for a chip. No collar, no tattoo, no chip. It suc
What GAO Found The National Nuclear Security Administration (NNSA)—a separately organized agency within the Department of Energy (DOE)—and its eight contractor-managed and -operated sites engage in Strategic Partnership Projects (SPP). These projects allow NNSA sites to perform work for other federal agencies and nonfederal entities and for those entities to benefit from the significant public inv
Hmm, I really need to write an article explaining how predictive surveillance likely works (behavioural prediction). Audience analytics → behavioural segmentation → propensity modelling → trajectory prediction → persistent identity → predictive surveillance Audience segmentation systems (like my toy one https:// lnkd.in/e995AzE4 ) *already* have the data for this. It's establishing groups (segment
Today, TIME convenes the third annual TIME100 Health Leadership Forum in New York City, featuring conversations on equity, longevity, and solution-driven care with leaders taking action toward a world with more reliable healthcare solutions. The TIME100 Health Leadership Forum will bring together individuals from the TIME and TIME100 Health communities and beyond who are actively shaping the healt
📺🎙️ Researchers just proved LG smart TVs can record room audio through the microphone even when the TV looks completely off, screen dark, unplugged from the internet entirely. It stores the audio locally, then quietly uploads it the moment the connection comes back. On top of that, these TVs constantly scan your home network, cataloging every phone, laptop, and smart device connected to your WiF
> The unfriendly skies: On United, your checked bags might be riding alongside some tank parts bound for Israel. > >United Airlines is set to resume the shipment of military hardware to Israel aboard passenger flights, according to shipping data reviewed by The Intercept. The airline, one of the biggest in the world, had halted the practice more than six months ago amid the outbreak of the war on
Magnus Hjort, director general of Sweden’s Psychological Defence Agency, said Russia is prioritizing information operations around elections in larger European countries, including upcoming German regional and local votes and France’s 2027 presidential race [politico.eu#1][nv.ua#1][digi24.ro#1][err.ee#1]. Hjort said his agency does not believe Moscow is targeting Sweden’s September 13 parliamentar
A measure of producer price inflation showed renewed pressure from rising energy prices last month, potentially adding to the case for an interest-rate hike at the Federal Reserve’s meeting next week. The producer price index rose 0.4% in August from the prior month — the most since May — and 5.4% from a year earlier, according to Bureau of Labor Statistics data out Thursday. Excluding food and en
A girl I went to college with told me a horrifying incident on Sunday. A trainer from her last gym tried to repeatedly hit on her. She simply left to join another place. He stalked her Instagram account. Downloaded pics. You know where this is going. Made nasty pictures and tried to blackmail her with them. The good part is her brother in law happens to be attorney and had dealt with cyber crimes
I just lost my job today, so here we go.. I'm a ~30 years Sysadmin/DevOps/Server Shepard/Graybeard whatever its called now UNIX nerd. Looking for work doing UNIX'y stuff. I mostly do automation, and have avoided working with AI and Kubernetes (though I have a bit of k8s experience at two jobs, I just prefer "the old way") Chef, Puppet, Cfengine, Ansible, Shell scripts, etc. I've built Linux distri
---------------- 🛠️ Tool: darknet-mcp-server =================== darknet-mcp-server is a new Model Context Protocol (MCP) server designed to unify dark web and threat intelligence sources for AI agents. It consolidates 16 different data sources into a single interface, exposing 66 distinct tools. This approach directly addresses the fragmented nature of dark web intelligence gathering, where anal
---------------- 🛠️ Tool =================== AI Second Brain is a free, open-source knowledge management system that pairs Obsidian with a Claude Code agent to build and maintain a personal wiki in plain markdown. The repo includes a vault template, agent skills, commands, and setup scripts. 🔹 Key Features • Agent-maintained wiki: An AI agent handles ingest, extraction, linking, and linting of r
today I created a burner google account. these days google requires an Android phone in order to even create an account so I used an old phone to scan the QR code ( I am pretty sure this lets google grab all of the metadata of the mobile phone ). after about 1 hour creating the account. google's automated system disabled the account however I appealed and got the account back. before someone asks
🚨 Surfshark discloses security incident after internal test server was exposed to the internet Surfshark says human error resulted in an internal engineering test server being misconfigured and publicly reachable, allowing an unauthorized third party to gain access. ⠀ The exposed environment contained limited internal engineering material, including parts of system binaries and configurations for
‘… force savings and shift investment from non-tradable to tradable sectors, including through exchange-rate management … periodic oversupply and imbalances is seen as an acceptable cost of building scale in strategic sectors. ‘The current-account surplus is … a means to achieve market and technological dominance. ‘… high savings, capital controls, exchange-rate management, and industrial policy …
Latest tariff scam warning: UPS and FedEx are retroactively going after purchases imported, not notifying you of additional tariffs, and sending to collections. I purchased a microfiber/microplastics filter for my washing machine. December of last year. I got a sketchy phishing email today claiming to be a debt collector. So I called FedEx. After 6 attempts navigating their automated fucking night
🚨🇫🇷 WiziShop and Dropizi dataset containing 511K+ records allegedly leaked on a cybercrime forum ⠀ WiziShop and Dropizi are French e-commerce platforms that help businesses create and manage online stores, with Dropizi focused on dropshipping. ⠀ A cybercrime forum actor using the handle ChimeraZ claims to have leaked a dataset associated with stores operating through WiziShop and Dropizi, conta
‘… organised syndicates packing stolen cars into shipping containers, sending them to intermediary ports to bypass detection before forwarding them to sanctioned nations. ‘… two-thirds of vehicles stolen in Melbourne are never recovered locally and that detections of Australian-registered vehicle parts overseas have increased. ‘Throughout the year, Victorian and Queensland police have raided shipp
‘To her surprise, Australia Post dropped several of those charges. ‘One email from Australia Post said: "I have reviewed the scans, and I can see two different sets of dimensions and weight recorded … I believe another parcel has moved closer to your [sic] causing the incorrect reading." ‘Although Australia Post uses an automatic size and weight scanning technology, which it claims is "highly accu
Fuck it, updated. Here's my list: - Rad debugger, Ryan Fleury - Xlibre - Neovim, Justin m keyes - Omarchy, David Heinemeier Hansson, white supremacist (wrote about culling immigrants and unhoused people, comparing it to shooting wolves that kill sheep) - Mullvad VPN, "Daniel Berntsson, co-founder and co-owner of Mullvad VPN, has donated 5 million Swedish kronor (around $514,000) of his personal fu
Parul Gulati advocates for a dynamic posting schedule on Instagram, featuring three posts daily for optimal business outreach. Morning content focuses on direct sales opportunities, catering to early risers. Afternoon visuals are crafted to captivate users as they browse. Finally, evening posts aim to activate laughter and relaxation, helping followers to decompress after a busy day, thereby impro
arXiv:2609.10594v1 Announce Type: new Abstract: Expert evaluation of jailbreak responses is costly and difficult to scale, so the community increasingly relies on automated evaluators to determine whether an attack succeeds. However, jailbreak studies typically validate their chosen evaluator independently, repeatedly spending resources on similar evaluation efforts while making results across pap
arXiv:2609.10601v1 Announce Type: new Abstract: Compound AI pipelines chain LLM calls, retrievers, and tools and are nondeterministic: sampling, model updates, and volatile tool responses make one input yield different outputs across runs. Such pipelines increasingly run across edge, cloud, and orbital nodes owned by no single party, whose optimizations discard intermediate results before inspecti
arXiv:2609.10608v1 Announce Type: new Abstract: Diffusion models have achieved remarkable success in generative tasks across various areas, however their training process raises significant privacy concerns, particularly under membership inference attacks (MIAs). Prior studies on privacy-preserving of diffusion models fail to balance privacy, utility, and efficiency. To address this gap, we propos
arXiv:2609.10611v1 Announce Type: new Abstract: Membership inference attacks (MIAs) have emerged as critical tools for auditing privacy risks in large language models (LLMs), aiming to determine whether a given text was included in a model's training corpus. However, most existing MIAs require access to per-token logits or probabilities, making them inapplicable in practice to proprietary LLMs tha
arXiv:2609.10612v1 Announce Type: new Abstract: Personalized learning systems rely on real learner data, including performance, behavior, and demographic information, but these data are highly privacy-sensitive. Differentially private (DP) synthetic data can support system development and educational research while reducing exposure of individual learners. Existing evaluations, however, assess pri
arXiv:2609.10613v1 Announce Type: new Abstract: In-context learning (ICL) jailbreaks reveal a critical vulnerability in multimodal large language models (MLLMs): harmful demonstrations in the prompt can induce unsafe outputs without modifying model parameters. Despite extensive empirical evidence, existing work lacks a principled understanding of why such jailbreaks reliably succeed or how their e
arXiv:2609.10627v1 Announce Type: new Abstract: Machine learning explanations reveal model behavior beyond predictions, creating attack surfaces for model confidentiality and data privacy. We systematize 25 studies that exploit explanations for model extraction, membership inference, and model inversion, treating attribute inference as partial inversion. Existing work is often labeled only black-
arXiv:2609.10631v1 Announce Type: new Abstract: A false data injection attack (FDIA) can change the estimated grid state while evading a residual-based bad data detector (BDD). Existing blind attacks learn a low-rank measurement subspace, but this algebraic view does not state the physical grid constraints that make an attack stealthy or the minimum information needed to recover the complete attac
arXiv:2609.10634v1 Announce Type: new Abstract: Trusted Execution Environments (TEEs) protect enclave memory from untrusted software but remain vulnerable to cache-based side-channel attacks due to shared microarchitectural resources. Existing countermeasures use techniques such as cache partitioning or randomization: these solutions are not ideal if a designer wants fine-grained configurations an
arXiv:2609.10744v1 Announce Type: new Abstract: Autoregressive language models can be used to transform a payload text into a stegotext of identical token length by preserving per-position rank information across contexts - a methodology we formalize as Contextual Autoregressive Rank Transcoding Steganography (CARTS). While the Calgacus construction of Norelli et al. demonstrated this phenomenon e
arXiv:2609.10762v1 Announce Type: new Abstract: Advances in large language models (LLMs) fuel the quest for scalable methods to assess the security of generated and security-sensitive software. Static analysis is widely adopted as a scalable, reproducible, and inexpensive security gate, but cannot directly observe runtime exploit behaviour. Vulnerabilities dependent on adversarial inputs, executio
arXiv:2609.10780v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly applied to penetration testing, but we still know little about what they can do or how they fail. We compare two PentestGPT-based systems: a legacy human-in-the-loop system running the open-weight Kimi K2.5, and a newer autonomous system running Claude Opus 4.8. Across three public targets, the auton
arXiv:2609.10854v1 Announce Type: new Abstract: Conventional vulnerability analysis relies on either system access or dynamic interaction, all of which may be unavailable to third-party analysts auditing closed-source, remotely hosted, critical in situ systems, or commercially gated software. Therefore, we propose a new paradigm of no-box vulnerability analysis in which neither access nor runtime
arXiv:2609.10871v1 Announce Type: new Abstract: The Agent2Agent (A2A) protocol, now governed by the Linux Foundation, is an open standard that enables autonomous AI agents to discover, authenticate with, and delegate tasks to one another across organizational boundaries. Designed to complement the Model Context Protocol (MCP) for tool integration, A2A is rapidly emerging as the horizontal communic
arXiv:2609.10892v1 Announce Type: new Abstract: When an indirect prompt injection succeeds against an LLM agent, the compromise is visible in the agent's own behavior: a benign prefix of tool calls, a poisoned observation, and a suffix of actions that serve the attacker. An operator needs three facts: where the attack entered, which steps it corrupted, and whether apparent poison was resisted. Exi
arXiv:2609.10935v1 Announce Type: new Abstract: Membership inference attacks (MIAs) try to determine whether a specific record was used to train a model, a privacy risk that matters in natural language processing (NLP), where training data can contain sensitive user text. This paper presents a controlled benchmark of membership inference vulnerability for text classification on the GLUE SST-2 sent
arXiv:2609.10952v1 Announce Type: new Abstract: Data poisoning corrupts training data to degrade a model or to plant attacker-controlled behavior. This study evaluates two representative training-time attacks, label flipping and backdoor poisoning, on MNIST and Fashion-MNIST with three baseline classifiers: Logistic Regression, Linear SVM, and Random Forest. Clean training is compared with poisoni
arXiv:2609.11008v1 Announce Type: new Abstract: Decentralized Finance (DeFi) has emerged as a rapidly growing blockchain-based financial service, where market transaction dynamics and underlying smart contract logic are intricately intertwined. This autonomous interplay, while eliminating centralized intermediaries, significantly expands the vulnerability surface of DeFi protocols to Price Manipul
arXiv:2609.11024v1 Announce Type: new Abstract: Autonomous agents increasingly perform long-horizon tasks involving tool use, persistent state, and consequential actions, raising a fundamental question: \emph{under what conditions does an agent cross the boundary of authorized execution while pursuing a legitimate task?} Existing studies often attribute such failures to adversarial instructions, m
arXiv:2609.11028v1 Announce Type: new Abstract: LM-agent benchmarks increasingly function as interactive evaluation infrastructure. Agents observe state, call tools, modify workspaces, submit artifacts, and receive rewards from outcome procedures. This interactivity makes evaluations vulnerable to reward hacking: an agent improves its measured score by exploiting the reward-relevant trajectory ins
arXiv:2609.11082v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) can ground large language model (LLM) outputs in external evidence, but it also exposes the system to knowledge poisoning. Representative attacks use multiple injected documents or templates that directly assert a target answer. We present ToxicRAG, a one-document-per-target attack that expresses misinformation as
arXiv:2609.11137v1 Announce Type: new Abstract: In February 2024 the U.S. Federal Communications Commission (FCC) placed AI-generated voices under the Telephone Consumer Protection Act (TCPA). Yet no peer-reviewed measurement says how much unwanted call traffic is placed by a machine, or how much of that machine speech is synthesized rather than played from a recording. We report both with a discl
arXiv:2609.11218v1 Announce Type: new Abstract: Web measurement studies rely on domain datasets such as Tranco to quantify the prevalence and impact of security issues at scale, but exhaustively analyzing these datasets is often infeasible because of the cost of advanced analysis techniques, requiring the use of sampling. Despite its widespread use, sampling remains largely guided by convention---
arXiv:2609.11251v1 Announce Type: new Abstract: Blockchains usually pay for authenticated reads by maintaining a structure that spans the entire state. We show how validators can support historical membership and exclusion proofs by authenticating each block's writes instead. A Block Update Digest (BUD) commits a write log whose predecessor pointers link successive modifications of each key. A Sup
arXiv:2609.11404v1 Announce Type: new Abstract: This paper presents DF-CAPTCHA, an active defense against real-time deepfake impersonation in voice and video calls. Instead of passively searching for artifacts, DF-CAPTCHA prompts the caller to perform simple challenge-response tasks that are easy for humans but difficult for current real-time deepfake systems to generate convincingly. The framewor
arXiv:2609.11411v1 Announce Type: new Abstract: Trusted Execution Environments (TEEs) have become a cornerstone of modern cloud computing, providing strong confidentiality and integrity guarantees for both code and data. A critical component of this trust model is remote attestation, which enables external entities to verify the authenticity and integrity of code executing within a TEE through cry
arXiv:2609.11413v1 Announce Type: new Abstract: Solana is a rapidly growing high-throughput blockchain platform that has attracted substantial liquidity and user activity. However, this expansion has also drawn the attention of illicit actors, who frequently leverage cross-chain bridges to route illicit funds onto Solana to obfuscate transaction lineage. Unlike EVM-compatible platforms, Solana fea
arXiv:2609.11442v1 Announce Type: new Abstract: Static side-channel attacks, which exploit halted-clock conditions to extract sensitive information, pose an increasing threat to chip security. To counter these attacks, various defenses have been proposed that monitor for abnormal clock behavior and trigger the clearing of sensitive data when clock anomalies are detected. In this work, we demonstra
arXiv:2609.11552v1 Announce Type: new Abstract: Certificate Transparency (CT) aims to reduce the trust required in Certificate Authorities (CAs) within the TLS certificate ecosystem. It is supported by all major browsers. The protocol obliges all CAs to record the certificates they issue in a public log, which itself is monitored for compliance and consistency by third parties. Given this complex
arXiv:2609.11757v1 Announce Type: new Abstract: Software agents are beginning to shop and pay on a person's behalf. Agent payment protocols such as AP2 produce cryptographically valid signatures for completed purchases, yet do not constrain the decisions that lead to them. Consequently, ordinary product-description text can steer a shopping agent into forming a cart that passes every protocol chec
arXiv:2609.11841v1 Announce Type: new Abstract: Semantic search is a core primitive of modern applications, powering recommender systems, web search, and retrieval-augmented generation for language models. The provider controls the index and query execution, leaving clients to trust that results come from the right algorithm over the intended index. A provider may truncate search to cut cost, bias
arXiv:2609.10543v1 Announce Type: cross Abstract: Privacy-preserving exchange designs price on a coarsened view of order flow. We show that a market maker committed to informationally efficient (posterior-mean) pricing on a signal strictly coarser than the flow it settles necessarily cedes a closed-form welfare transfer to traders -- the privacy subsidy -- and that no rule restricted to the coarse
arXiv:2609.10830v1 Announce Type: cross Abstract: When a language model finds a sentence unusually cheap to predict, it is tempting to conclude that the sentence was in its training data. Almost every published test of that inference has had to guess which sentences were in the training data, the members, and which were not. This paper removes the guessing. Two model families, OLMo-2 and Pythia, p
arXiv:2609.10877v1 Announce Type: cross Abstract: This paper studies fundamental graph optimization problems under differential privacy (DP) and shows new, reconstruction-based lower bounds. We consider a graph $G = (V, E, \vec{w})$ where the vertex set $V$ and edges $E$ are public and the weights $\mathbf{w}:E\rightarrow \mathbb{R}$ must be kept differentially private under an $\ell_1$ neighborin
arXiv:2609.10992v1 Announce Type: cross Abstract: The integration of Large Language Models into daily tasks relies on context-rich instructions, inevitably exposing sensitive user information. Current privacy-preserving methods typically employ context-agnostic static rules, causing severe utility degradation. However, the specific mechanisms governing how sanitization impacts downstream performan
arXiv:2603.22437v2 Announce Type: replace Abstract: We present mmFHE, the first system that executes the entire cloud-side mmWave sensing pipeline including the DSP and ML inference under fully homomorphic encryption (FHE). mmFHE encrypts range profiles on an edge device after lightweight plaintext preprocessing and executes the entire mmWave signal-processing and ML inference pipeline homomorphic
arXiv:2606.04990v5 Announce Type: replace Abstract: Large language model (LLM)-based agents are evolving from passive text generators into autonomous systems capable of planning, tool use, retrieval, memory access, environmental interaction, and multi-agent collaboration. These capabilities expand agent autonomy, but also make agent behavior harder to verify, debug, and audit. Final-answer accurac
arXiv:2506.13160v2 Announce Type: replace-cross Abstract: Deep neural networks (DNNs) rely heavily on high-quality open-source datasets (e.g., ImageNet) for their success, making dataset ownership verification (DOV) crucial for protecting public dataset copyrights. In this paper, we find existing DOV methods (implicitly) assume that the verification process is faithful, where the suspicious model
arXiv:2609.11614v1 Announce Type: new Abstract: Classical market-making strategies based on stochastic control, such as the Avellaneda-Stoikov and the Gu\'{e}ant-Lehalle-Fernandez-Tapia (GLFT) extension, provide closed-form quoting rules, but rest on assumptions that break down at realistic microstructure timescales. One of them is that order flow is stationary, while empirical evidence points to
arXiv:2609.11905v1 Announce Type: new Abstract: We develop Entropic Value-at-Risk (EVaR) parity for tempered stable returns. EVaR-based inverse risk parity (IRP) and equal risk contribution (ERC) portfolios are constructed using multivariate normal tempered stable models and independent component analysis with tempered stable components. We derive the corresponding asset-level EVaR and EVaR-deviat
arXiv:2609.10587v1 Announce Type: cross Abstract: We develop an expected shortfall factor model (ESFM) to estimate and price common variation in the severity of lower-tail losses in large panels of asset returns. Mean factor models describe common variation in average returns, while quantile factor models describe common movements in tail thresholds. ESFM instead captures common variation in the a
arXiv:2609.10865v1 Announce Type: cross Abstract: Constructing a confidence interval for the difference between two independent binomial proportions involves a nuisance direction that is not identified by the estimand. The one-sample Wilson score interval inverts a scalar score test, but has no direct bivariate analogue isolating the difference: inverting the joint normal approximation yields an e
arXiv:2609.11575v1 Announce Type: cross Abstract: Modeling the joint distribution of extreme values in high-dimensional financial time series is challenging because extremes are sparse and locally extreme observations are not necessarily extreme relative to their full marginal distribution. To address this, we introduce a time-dependent network H\"usler-Reiss model in which market-informed adjacen
arXiv:2606.13992v2 Announce Type: replace Abstract: We develop an Affine Holonomy Group Quantization framework for the Heston stochastic volatility model. The Heston affine pricing symbol is decomposed into a finite symplectic quadratic sector and a complementary holonomy sector, leading to a Poincare-Cartan form whose characteristic flow yields the affine Riccati dynamics. Momentum polarization g
arXiv:2606.20041v2 Announce Type: replace Abstract: We propose an AI economist agent for economic and financial scenario analysis. Scenario design often requires analysts to assess emerging risks with limited historical precedent, combine information from many sources, and translate qualitative mechanisms into internally consistent quantitative paths. Large language models (LLMs) can search and sy
arXiv:2606.14887v2 Announce Type: replace-cross Abstract: Models whose predictions depend on only a handful of well-constrained parameter combinations, termed sloppy models, are ubiquitous in nonlinear stochastic systems. The information-geometric approach to sloppiness advocates using the symmetrized Kullback--Leibler divergence and its associated Hessian, the Fisher Information Matrix (FIM), as
Ukraine hit targets across four Russian regions overnight with Saratov strikes being the most confirmed and dramatic. Ozon's massive logistics hub in Saratov (100,000+ m², 30M+ items stored, up to 900K orders/day). The Saratov oil refinery (Rosneft, repeatedly targeted before) also reportedly caught fire. In Volgograd, the governor confirmed debris hit an industrial facility and an apartment build
Widespread damage across Ni’ihau, Kaua’i and O’ahu after reportedly first-ever recorded tornado watch in Hawaii Hurricane Lowell caused significant damage to Hawaii’s western islands this week. On Tuesday night the centre of Lowell tracked 50 miles (80km) from the island of Ni’ihau in Kauai County as a category 2 storm with winds of up to 110mph. The strong winds resulted in widespread power outag
"This is history, this is history." That's what photojournalist Spencer Platt was thinking 25 years ago as he photographed the events unfolding on the morning of 9/11. Why it matters: Platt and fellow Getty photographer Mario Tama told Axios what it was like on the ground that day. Trained as photojournalists, their instinct was to move toward the danger to document history. Tama had just been hir
I don't have a lot of reach on here, so boosts would be greatly appreciated. I'm looking for someone to join my team at the Lean FRO as a metaprogrammer with a view towards library quality. I'm working toward a future in which ITPs are a joy to use productively for mathematics and software verification. I believe that to get there, we need to get a lot better at building libraries that are complet
Two weeks after the devastating flash floods, the number of dead and missing continues to rise. At Kathmandu’s public hospital, as well as outside the military base in Bidur, hundreds of photographs of missing people have been put up. Hundreds of families come forward, hoping for news – good or bad – of one or more loved ones. A race against time is underway: helping survivors, finding those still
Perhaps the lizard species in the world that has the most well-studied behaviour and communication in the wild is the Iberian wall lizard, Podarcis hispanicus. They don't look particularly exceptional - and possibly they aren't, we just haven't studied other lizards as much. Their social life is fairly typical for lizards: Both males and females are mostly solitary and territorial, though females
Ten years ago today, I went on my first date with Meg. We'd been chatting online for a few weeks, but it was the first time we'd met in person. Meg told me I stood out because I was the only man who used *paragraphs* in his first message and actually responded to her profile instead of a canned response. A low bar, but I cleared it, apparently. I usually planned dates to be open-ended, starting wi
As we enter the fall mountain biking season, riding at night is part of the story now. A few notes. First, due to the summer heat, night riding was a thing in July and August. That's new. Second, I'm convinced more than ever that night riding is a great way to not only enjoy your hobby but also to get better. I'd argue one can get significantly better spending time riding at night. My theory is si
🚨🇦🇪 ECC Group database containing 52K+ records allegedly leaked by AnkaTeam ⠀ ECC Group is a Dubai-based group of companies operating across construction, engineering, fit-out, manufacturing, building materials, facilities, and real estate development. ⠀ An actor using the handle SALDIRGAN, identifying with AnkaTeam, claims to have leaked a database associated with ECC Group’s official website.
Yeah, I think I can summarize it as this: Javelins _frontload_ the value. Even a complex javelin breakdown you are going to get a lot of yield very quickly in the beginning. There's some nuance to this and you _need_ good equipment to do it, but you can do it systematically and usually strip out most of the critical systems within a single shit. Even when the configuration is slightly weird, you a
Tearing my hair out now with Debian Trixie. Working on some ansible to build a proxmox server... during the play I'm switching from a DHCP assigned address used while the device is imaged via PXE, to a static address... and at this time I'm trying to set nameservers in resolv.conf. DHCPCD seems to be being triggered by something in the background, and it keeps overwriting the resolv.conf I've adde
The taste is unlike anything I've tried, but it certainly makes me think of the forest. It's sweet and fruity, but with a little bitter aftertaste, similar to that of lingonberries. The color is a dark red-orange when made following the recipe, and a pink-orange when diluted as in the photo. Process-wise, it's important to freeze the berries first to break down the starch into sugar, and to boil i
>Bin Laden’s plan was to trap the U.S. in expensive, foreign quagmire wars that would trigger a domestic economic collapse. It was patterned on the strategy that the local mujahideen, he and other foreign fighters, and the U.S. and its allies had used to defeat the Soviet Union in Afghanistan. > >Not only does bin Laden’s strategy continue apace long after his 2011 death, but it also spawned an ev
ESFOTX A slow moving storm system will bring 12 to 15 hours of rain to Central Washington Saturday night through Sunday night. Potential rain amounts will vary from half an inch to up to two inches. This will bring an elevated risk for debris flows over sensitive burn scars within steep terrain as well as isolated rock slides. The potential hazard threat will increase slowly throughout the event a
Solicitar senha. Solicitar token. Token inválido. Aguardar. That's the full operator menu for VX-Pack — a Brazilian-origin AiTM phishing-as-a-service kit targeting banks in Brazil and Portugal. Request password. Request token. Invalid token (ask again). Wait. One operator, one victim, one browser, in real time. Nearly every AiTM kit — Evilginx, Tycoon 2FA, EvilProxy — is a reverse proxy. It silent
🔹 Latest Bulletins CVE-2026-89332 - Kiro IDE Sensitive Workspace Data Exfiltration via Agent-Written Workspace Configuration Bulletin ID: 2026-111-AWS Scope: AWS Content Type: Important (requires attention) Publication Date: 09/11/2026 12:00 PM PDT Description: Kiro IDE is an agentic development environment that makes it easy for developers to ship real engineering work with the help of AI agents
For anyone following this # selfhosting thread, I am already in love with Immich: Trivial integration with Authentik for OAuth SSO hierarchical albums Effortless launching as an app on TrueNAS tagging Customizable share links. (E.g. if I want the public link to my album to be /s/Christmas2025, I can do that) It does not upload folders (ente did that trivially), but I still like it. Ente is about p
Is anyone else in a panic over the Gemini 2.5 model generation (Pro, Flash) being sunset in October before there's even any Pro class model in general availability (with geo restrictions etc.)? Google wants everyone to migrate to 3.x Flash, which beats the older Pro models on the benchmarked tasks, but isn't the same thing as the Pro class on reasoning-heavy tasks like complex reasoning on very la
RE: https:// mathstodon.xyz/@tao/1172373221 60500501 I'm somewhere in the middle on the LLM thing. I definitely use LLMs and enjoy using them to accomplish things I wouldn't otherwise have the time to do. I'm not a doomer - I'm also not a AGI fanatic. I think it's a useful tool. That said, I think it _is_ true that AI use so far is that it has a negative effect on human relationships. Whether it's
Stressed UK homeowners tell of cracks opening up in the walls around them after two record-hot summers “We’ve got one wall with an absolutely huge diagonal crack, through which we can see through to the other room; you could easily put your fingers in it and chunks of plaster are falling off,” says Henry Biggs, 49, from Stroud in Gloucestershire. He and his partner started noticing cracks in their
I built a browser app that turns a prompt into a parametric 3D model. You steer it by adding reference images or drawing directly on a render of the model. The LLM then looks at a render of what it just built and fixes its own mistakes (usually converges in 2-3 rounds). I find it quite satisfying to watch the model iterate on its own work. No installation and no account are required. Everything st
As trust in institutions wavers, experts say local, collective action is more effective in driving change Wildfires, excess deaths, hosepipe bans, school closures: this summer many British people have experienced the often devastating effects of the climate crisis in ways they had not before. In that sense, policymakers and climate experts have a unique opportunity to engage with the public about
Delaying this week's blog article, I’ve been writing a few articles that are out there even for me: high-dimensional metadata entity partitioning. Basically, audience segmentation...but for EVERYTHING. I already explored the concept in my cross-modal RAG series https:// mostlylucid.net/blog/doomsumma rizer-deep-research (culminated in DoomSummarizer). This is the generalisation of that pattern eve
📡✊ SpaceX just abandoned its plan to throttle Starlink speeds for high-usage residential and satellite customers, walking it back entirely after widespread backlash and mass customer appeals. Some subscribers had already seen their speeds silently dropped to around 10 Mbps, despite paying for "unlimited" plans, before the public pressure forced SpaceX to reverse course. This isn't even the first
backflow versus drain over a little time. Interactive applications notice this first because they depend on low and consistent latency: voice/video calls, remote shells, streaming, games, and other real-time workloads immediately expose a 100–700 ms stall as frozen audio, lag, jitter, delayed input, or dropped frames. Less interactive workloads such as downloads, background sync, mail retrieval, b
We shared a one-off share of over 400 compromised PaperCut NG/MF instances (via CVE-2026-81578/CVE-2026-82078) observed by @ greynoise . IP data in our Compromised Website reporting for 2026-09-11, tagged 'papercut-compromise'. Dashboard Tree Map stats: https:// dashboard.shadowserver.org/sta tistics/combined/tree/?date_range=other_value&day=2026-09-11&source=compromised_iot&source=compromised_web
🚨🇫🇷 Ekolis data leak allegedly exposes driver, vehicle, and logistics information ⠀ Ekolis is a French company specializing in telematics and connected-device solutions for the trucking and logistics industry. ⠀ An underground forum member using the handle 4me44 claims to have leaked data associated with Ekolis following an alleged breach on September 5, 2026. ⠀ Claimed exposed data includes: ⠀
// # WindingAroundHer # writing 8. self-aware Only by denying her own self can a mouse believe the system-of-Me is not defective. And so the mouse-I invents tricks to lie to herself and pretend that the turn is not there. They keep the robot’s eyes locked forward, as robot-me walks in a straight line with forced unawareness of the curve toward from a self I is not allowed to imagine. As the robot
🚨🇦🇺 TIC Group source code allegedly leaked following September breach ⠀ TIC Group is an Australian supply chain and retail logistics provider offering reverse logistics, product recovery, warehousing, and related services to retailers and brands. ⠀ A forum moderator using the handle 888 claims to have published source code allegedly stolen from TIC Group during a breach in September 2026. ⠀ Cla
// # WindingAroundHer # writing *one of these days I will get "posting as a reply" right and not need to fix it three times 😅* 8. self-aware Only by denying her own self can a mouse believe the system-of-Me is not defective. And so the mouse-I invents tricks to lie to herself and pretend that the turn is not there. They keep the robot’s eyes locked forward, as robot-me walks in a straight line wi
I finally fixed my doorbell. The internal battery failed, and so when the "ring the internal chime" function was enabled, the device would die because it had no charge to function while it shorted out the contacts. Turns out it was a well known issue, and of course the entire unit was designed to be replaced, not repaired. Classic planned obsolescence. I need to do better/more research before I bu
A real-life OpSec story on fingerprinting and pattern disruption, that happened to me few days ago, and I wanna share it with you! The other day, I was walking down the street, 100% alert to my surroundings. A woman walked right past me, and I didn't recognize her at all... Until she called my name and reached out for a handshake. Turns out, she was the director of my previous corporate job. My br
"Mamdani, meanwhile, said voters in the Bronx and Queens told him they were drawn to Trump for 'two major reasons'. 'One was that they wanted an end to forever wars. They wanted an end to the taxpayer dollars we had funding violations of human rights," Mamdani said. "And they wanted to address the cost-of-living crisis'." This phenomenon COULD be strategically significant for the DSA insurgency ag
Sometimes my wife gets very annoyed at me because no matter how hard I try erase some parts of my Singapore upbringing, it does peek through very often when it comes to matters where I usually have a lot of comfort and privilege there. It’s hard to describe. I think that they’ve largely removed most elements of physical discomfort for the middle and upper class of citizens and permanent residents
Astronomers are developing observational technologies to study distant planetary systems, including the rocky exoplanet LHS 1140b, which orbits a red dwarf about 48 light-years from Earth. Recent observations detected helium escaping from its upper atmosphere; heavier compounds such as water could remain lower down, but the available evidence does not establish that the planet supports life or has
I might have stumbled into partial string matching out of necessity. I am dealing with English text, mainly ASCII set (can be converted to handle Unicode/UTF-8 as supported by Python v3.12). (In both scenarios below characters matching "[^A-Za-z0-9]" are removed; space is also removed for the first case; it is used as splitter to form set of "word"s for the second case.) Given 2 strings: - first,
Thing 2 (the one that annoys me): leaving apostrophes out because you don't have any idea how they work and you know they might creep in at inappropriate times because of that. In the memes attached here we see possessives without appostrophes and (I think this is more recent) even without the "s": "A homeless man sign said" "...what The Rock end goal is." The second tweet (or whatever it is) is e
Ukraine Brief — 2026-09-13 Ukraine's confirmed 3,000-km drone reach to Russia's Yamalo-Nenets gas heartland and Russia's third ballistic-missile strike against Zaporizhstal will very likely sustain mutual infrastructure escalation through October. Ukraine will very likely hit at least two more Russian energy facilities before October 31; Russia will very likely strike Ukrainian industrial infrastr
A conflict no longer has to spread by armies crossing the next border. It can spread through weapons, drones, mercenaries, terrorist networks, refugees, disrupted trade routes and the intervention of outside powers. One war can provide the weapons for another, weaken the neighbouring state, draw in another patron and create the conditions for the next conflict. Seen on a map, the scale is disturbi
Compounding all of this is the self-inflicted retreat of the United States from its role as the principal guarantor of the international order. Whatever one thinks of the way America exercised that role, its withdrawal leaves a power vacuum across precisely the regions where conflicts are beginning to overlap. No other power is currently capable of filling that vacuum, and several of those involve
Will OST just result in structurelessness? Jo Freeman's 1970 essay "The Tyranny of Structurelessness" comes up now and then as an argument against DP2. The reasoning runs like this. Freeman showed that "leaderless, structureless" groups don't actually eliminate power; they just hide it in informal, unaccountable elites. DP2 removes hierarchy. Therefore, DP2 is exactly the kind of structurelessness
🚨🇲🇽 UASLP student database containing 940 records and facial images allegedly leaked ⠀ Universidad Autónoma de San Luis Potosí (UASLP) is a public university in San Luis Potosí, Mexico, offering undergraduate, graduate, research, and professional education programs. ⠀ A forum actor using the handle zfo claims to have obtained and released a UASLP student dataset containing information on approx
Software intensification, the layering of abstraction and dependencies such that the inputs needed to accomplish the same effective productivity are greater, is accelerating. Even in the places without AI slop or big tech enshittification driving it, the toolchains To build software get more complex. To build librsvg I need ninja, meson, cmake and make, python and rust toolchains. I deeply appreci
🚨🇧🇪 Fédération francophone de Gymnastique dataset containing 209K+ member records allegedly breached ⠀ The Fédération francophone de Gymnastique is a Belgian sports federation that promotes and organizes gymnastics across French-speaking communities and affiliated clubs. ⠀ A forum actor using the handle Venus1337 claims to have obtained data associated with ffgym.be, along with access to an adm
I wish something like the Noise protocol specification ( https:// noiseprotocol.org/noise.html ) existed for PQNoise. The PQNoise paper is rightly heavily focused on proving the properties of their constructions, but as a result the information that practitioners need in order to implement pqnoise patterns is diffused throughout the paper, and is mostly expressed as diffs relative to classical Noi
My natural accent (i.e. not the posher one I automatically put on when I need people to take me seriously) is a mixture of a working-class Estuary/South-East England accent (where I was born and lived for 17 years) and Belfast (where I've lived for 22 years*). So a lot of my slang/idiolect is from Belfast, as is most of my syntax and the overall cadence/tone of how I speak. I have the singsong lil
I guess I'm probably the last human on earth to sign up for a Costco membership, but I did that today. And I had to have a whole fight with them about "no, I'm not installing an app." The woman working the registration desk was perfectly nice, but then simply would not budge. Eventually she asked "would you like to speak to my supervisor?" and I said yes. Then the supervisor said she needed to get
Italy’s government is finalizing its budget law and is considering a 5% flat tax on contractual pay increases for younger workers. Economy Minister Giancarlo Giorgetti said the proposal would require employers’ cooperation and that the coalition would discuss it in the coming weeks [repubblica.it#1][repubblica.it#3][iltempo.it#1]. Giorgetti also warned that higher interest rates are raising public
Recent U.S. releases of documents, photographs and videos about unidentified anomalous phenomena have renewed public discussion of possible non-human intelligence, but the material has not established that extraterrestrial craft or life exist. The Pentagon’s UAP caseload includes more than 2,000 reports dating to 1945, while some released images may have conventional explanations, including image
‘We study how wars and military threats affect financial markets. Leveraging more than 300,000 monthly price observations since 1822, we create an external currency bond index for more than 90 countries — the EXBI. Using the EXBI, we document large effects of wars on returns and borrowing costs. In a global external bond portfolio, a one-standard-deviation war shock lowers returns by five percenta
I wanted to know how supply chain malware research is done. Who better to chat to about this than Paul McCarty from OpenSourceMalware! We did a 85 minute deep dive into how malware research is conducted end-to-end, what Paul's seeing in his research and what that tells us about the tradecraft, and then we rounded the episode off with some practical advice for defenders. We covered how he collects
Two arXiv papers submitted or revised on September 11 examine how generative AI affects mathematics learning and how researchers should test AI tutoring systems. A living meta-analysis of 34 studies estimated that generative-AI interventions had a positive effect on learning, with a standardized effect size of g = 0.57 and a credible interval of 0.36 to 0.79. It found no evidence of publication bi
arXiv:2609.11952v1 Announce Type: new Abstract: Chemistry and materials agents integrate literature retrieval, candidate generation, property prediction, and protocol planning into continuous discovery workflows. Consequently, the relevant safety question is shifting from whether a model answers a hazardous question to whether an agent releases a hazardous protocol through a tool-mediated workflow
arXiv:2609.12001v1 Announce Type: new Abstract: Agent skill registries screen what they publish. OpenClaw's security team reported that its scanners overlap on at most 10.4% of combined positives, and 81.9% of flagged skills are caught by one scanner alone. We take that as given, and suggest the open question is not which scanner is right but which is being asked. Each answers a form of "is this s
arXiv:2609.12145v1 Announce Type: new Abstract: As the quantum computing field transitions toward Fault-Tolerant Quantum Computing (FTQC), intensive efforts are focused on scaling architectures and realizing active error correction. However, this shift introduces security surfaces that remain largely unexplored. Fault-tolerant quantum computers pair a quantum processor with a classical decoder tha
arXiv:2609.12263v1 Announce Type: new Abstract: Graph-based network intrusion detection systems (GIDS) report strong benchmark detection metrics, but those metrics establish little about deployability. We approach the problem from first principles: rather than inheriting the preprocessing, windowing, and thresholding conventions of each published system, we ask what a controlled comparison require
arXiv:2609.12315v1 Announce Type: new Abstract: Blockchain application attacks, targeting business logic bugs in decentralized applications (dApps), have been an increasing concern to their developers and users, causing significant financial loss. Existing attack detectors either rely on handcrafted rules for detection, or need difficult-to-obtain smart contract source code to analyze attack trans
arXiv:2609.12360v1 Announce Type: new Abstract: Critical-infrastructure knowledge is distributed across heterogeneous, incomplete, and weakly structured evidence, making dependency models difficult to construct automatically and difficult to trust. Large language models (LLMs) can extract structured knowledge from such evidence, but direct LLM-to-graph generation risks unsupported relationships, i
arXiv:2609.12378v1 Announce Type: new Abstract: Cloud LLM services typically require users to send prompts to a model provider, creating a privacy risk. Fully homomorphic encryption (FHE) lets a server perform inference without decrypting the input, but representing data as ciphertexts adds storage and computational overhead. In CKKS-based LLM inference, the packing scheme maps logical tensors to
arXiv:2609.12426v1 Announce Type: new Abstract: Insecure Direct Object Reference (IDOR), often modeled as Broken Object-Level Authorization (BOLA), remains prevalent in Java database applications because identity and authorization checks at the controller or service layer are disconnected from SQL execution based on resource identifiers. Existing work largely detects these vulnerabilities but offe
arXiv:2609.12450v1 Announce Type: new Abstract: The security and liveness of Proof-of-Work (PoW) blockchains fundamentally depend on the economic rationality of miners. Existing incentive-driven denial-of-service attacks, such as BDoS, can deter rational miners from participating, but require the attacker to continuously absorb substantial economic losses and are therefore difficult to sustain in
arXiv:2609.12488v1 Announce Type: new Abstract: Commercial security gateways repeatedly ship implementation bugs in the code path between the network and the policy decision: hand-written enforcement logic that diverges from the policy author's intent, and ad-hoc request parsers at the network boundary that introduce memory-safety flaws of their own. In both cases the bug is in the deployed enforc
arXiv:2609.12554v1 Announce Type: new Abstract: Metaverse platforms increasingly derive their success from user-generated virtual worlds: self-contained social and interactive environments, which can be created by any ordinary user and scale to billions of visits. Platforms such as Roblox, Horizon Worlds, and VRChat now host millions of creator-built worlds that govern how users see, hear, and int
arXiv:2609.12571v1 Announce Type: new Abstract: Privacy impact assessment (PIA) is a critical instrument for institutions to proactively identify privacy risks and develop mitigation strategies before system deployment. While mandated across regulatory and institutional contexts, executing PIA requires extensive privacy and technical expertise, posing a particular challenge for teams without acces
arXiv:2609.12580v1 Announce Type: new Abstract: Arm TrustZone for Armv8-M isolates Secure and Non-secure software, but developers must still coordinate peripheral attribution, interrupt routing, initialization, and gateway interfaces across separately built firmware images. Inconsistent assumptions between these images can compile successfully and emerge only as faults on the target device. We pre
arXiv:2609.12624v1 Announce Type: new Abstract: With transciphering, a server converts symmetrically encrypted records into homomorphic ciphertexts without learning the records or the symmetric key. For the PASTA cipher, this conversion involves dense linear maps whose implementation depends on how record words are arranged in the ciphertext. We ask whether rearranging a fixed batch can reduce its
arXiv:2609.12646v1 Announce Type: new Abstract: As the threat landscape against industrial control systems is dynamic, effective security requires detection strategies capable of timely reactions to these evolving threat situations. To address this problem, we propose the idea of a counter-threat intelligence sharing based mechanism to select appropriate detectors for current circumstances. To hig
arXiv:2609.12727v1 Announce Type: new Abstract: Can a finite verifier obtain public, model-relative evidence about response latency for sequential computation? Verifiable delay functions (VDFs) make this possible in principle: evaluation requires T sequential steps, whereas verification is efficient in the security parameter and polylogarithmic in the numerical value of T for standard construction
arXiv:2609.12810v1 Announce Type: new Abstract: Cyberwarfare has become a key component of contemporary geopolitical conflict. However, there has been extremely limited systematic investigation into how cybersecurity is handled by military organizations and personnel. The military context is unique compared to other operational ones, with immense resource availability (U.S. military spending appro
arXiv:2609.12841v1 Announce Type: new Abstract: Mapping observed system behavior to standardized frameworks like MITRE ATT&CK is essential for threat-informed defense, but remains largely manual. Existing automated methods depend on Cyber Threat Intelligence reports, which offer only retrospective accounts of attacks. Low-level telemetry, i.e. kernel-level system calls, instead provides evidence o
arXiv:2609.12911v1 Announce Type: new Abstract: Hybrid fully homomorphic encryption~(FHE) inference improves the practicality of private inference by letting the server evaluate linear layers homomorphically while the client decrypts and applies nonlinearities. Recent schemes attempt to protect model confidentiality by returning noisy, output-permuted responses and appealing to shuffle-model diffe
arXiv:2609.13004v1 Announce Type: new Abstract: Cross-chain bridges move value between blockchains. Intent-based bridges are a variant where a solver fulfills a user's declared outcome and an off-chain settlement layer later reconciles the fill against the deposit. Existing smart-contract fuzzers and static analyzers only flag known-bad code patterns or require protocol-specific hand-written asser
arXiv:2609.12008v1 Announce Type: cross Abstract: Software security testing (SST) is essential for identifying vulnerabilities and improving software security, but developers often face practical challenges when selecting tools, configuring test environments, interpreting scanner outputs, testing authentication workflows, and acting on reported vulnerabilities. This study empirically characterizes
arXiv:2609.12341v1 Announce Type: cross Abstract: Authorship attribution models can re-identify users from seemingly anonymized text by exploiting stable stylistic fingerprints, even after explicit identifiers are removed, posing a growing privacy risk for text publishing and analytics. This risk extends to speech-derived text such as ASR transcripts of meetings and call-center conversations, wher
arXiv:2312.11581v3 Announce Type: replace Abstract: The pandemic in 2020 and 2021 had enormous economic and societal consequences, and studies show that contact tracing algorithms can be key in the early containment of the virus. While large strides have been made towards more effective contact tracing algorithms, we argue that privacy concerns currently hold deployment back. The essence of a cont
arXiv:2509.17836v2 Announce Type: replace Abstract: Machine Learning (ML) techniques have shown strong potential for network traffic analysis; however, their effectiveness depends on access to representative, up-to-date datasets, which is limited in cybersecurity due to privacy and data-sharing restrictions. To address this challenge, Federated Learning (FL) has recently emerged as a novel paradig
arXiv:2509.22215v3 Announce Type: replace Abstract: Security verification of communication protocols in industrial and safety-critical systems is challenging because implementations are often proprietary, accessible only as black boxes, and too complex for manual modeling. As a result, existing security testing approaches usually depend on incomplete test suites and/or require labor-intensive mode
arXiv:2604.05793v2 Announce Type: replace Abstract: In LLM agent pipelines, prompt privacy risk propagates beyond a single model call: raw user content enters retrieval queries, memory writes, tool arguments, OCR-derived text, and logs, and every downstream copy inherits what the first write contained. Existing de-identification pipelines protect document boundaries but not this cross-stage surfac
arXiv:2605.01301v2 Announce Type: replace Abstract: Collaborative perception allows connected and autonomous vehicles (CAVs) to improve perception by sharing sensory data, but it also introduces security risks from manipulated inputs. Prior work shows that attackers can spoof or remove objects by fabricating shared data, yet the practicality of such attacks in real-world driving remains unclear. E
arXiv:2608.00583v3 Announce Type: replace Abstract: Chain-of-thought (CoT) monitors are reported by their aggregate accuracy on a pool of reward hacks. We show that this number is a false average. On Terminal Wrench, about 77% of hacks are given away by the actions alone, and the monitor's pooled accuracy is dominated by them; on the remaining 23%, where the reasoning is the only signal, the same
arXiv:2608.02681v2 Announce Type: replace Abstract: Batch prompting is a practical inference strategy for large language models, but its safety implications remain underexplored. We show that the success of batch prompting for utility does not extend to safety: a harmful question that is reliably refused in isolation can elicit a harmful response when embedded in a batch of benign questions. We id
arXiv:2403.05864v3 Announce Type: replace-cross Abstract: In human-centric Cyber-Physical Systems (CPS), personalized Deep Reinforcement Learning (DRL) agents must share fine-grained control actions with cloud services, exposing sensitive private states to inference attacks by honest-but-curious adversaries. Static privacy models fail to address the dynamic nature of human interactions. This paper
arXiv:2509.11080v4 Announce Type: replace-cross Abstract: Recommender systems (RecSys) have been widely applied to various applications, including E-commerce, finance, healthcare, social media and have become increasingly influential in shaping user behavior and decision-making, highlighting their growing impact in various domains. However, recent studies have shown that RecSys are vulnerable to m
arXiv:2607.02304v2 Announce Type: replace-cross Abstract: We consider grassroots platforms -- distributed systems of agents consisting of people identified by self-chosen public keys and their machines (smartphones) -- and wish to make them secure against \emph{major faults}: the loss of their private keys and/or their smartphones. As grassroots platforms have no global resource to rely on for rec
arXiv:2609.12477v1 Announce Type: new Abstract: An ensemble of roughly 3,000 signals over 20 assets was reported to have approximately 90% correlation with the leading component of the asset-space return structure. Does having about 158 signals per available linear dimension explain that alignment? The population answer depends on the research process's design distribution and its relation to retu
arXiv:2609.12666v1 Announce Type: new Abstract: In this note, we propose a general structural approach to model a default time $\tau$ as the first-passage time (FPT) of a (``firm-value'') process $S$ below a (``debt'') barrier $K$ that comply with a pre-specified survival probability curve $G(t)=\Pr(\tau>t)$. Following an idea of Mbaye and Vrins (Mathematical Finance, 2022) applied to reduced-form
arXiv:2609.12878v1 Announce Type: new Abstract: The favorite--longshot bias (FLB) is one of the most persistent return patterns in betting markets, but its prevalence in modern prediction markets and its relation to recurrent longshot demand remain unclear. We study these questions in Polymarket, a large and increasingly important prediction market, using 588 million trades by 2.48 million account
arXiv:2609.12976v1 Announce Type: new Abstract: Whether medical technology reduces or increases demand for professional care is central to understanding its effects on healthcare utilization and costs. We study automated insulin delivery (AID) adoption among 1,608 adults with type 1 diabetes treated in four specialist clinics of the Italian National Health Service, 283 of whom adopt AID. Because a
this outlet adds: 12976v2, replace, sensi
arXiv:2609.13068v1 Announce Type: new Abstract: Empirical researchers often use binary classification trees to identify subgroups at risk of adverse outcomes. We compare two classification tree algorithms in this policy-targeting context: classical classification and regression trees (CART) and the maximizing-distance final-split approach (MDFS). We establish a theoretical setting in which MDFS id
arXiv:2609.12793v1 Announce Type: cross Abstract: Stock price prediction remains challenging due to the non-stationary and noisy nature of financial time series. Existing deep learning models often rely on rigid decision-level fusion, ad hoc hyperparameter tuning, and compressed final-layer outputs, causing information loss, overfitting, and limited cross-market generalization. We propose VertiFus
arXiv:2509.22985v2 Announce Type: replace Abstract: Liquidity withdrawal is a critical indicator of market fragility. In this project, I test a framework for forecasting liquidity withdrawal at the individual-stock level, ranging from less liquid stocks to highly liquid large-cap tickers, and evaluate the relative performance of competing model classes in predicting short-horizon order book stress
arXiv:2603.29530v2 Announce Type: replace-cross Abstract: This paper studies proportional risk sharing at claim occurrence time in community-based insurance. Each participant is modeled by an individual Cram\'er-Lundberg surplus process, and, whenever a claim is reported within the pool, its cost is redistributed according to a fixed allocation matrix. We compare the infinite-time ruin probability
arXiv:2605.18019v2 Announce Type: replace-cross Abstract: We propose a data-driven Fourier-trained neural-network method for estimating fixed-horizon probability densities from empirical characteristic-function (CF) information. The estimator is a positive Gaussian--Laplace mixture with closed-form CF, so training can be performed directly in Fourier space while preserving nonnegativity and unit m
What GAO Found The National Register of Historic Places is the nation’s official list of historic places worthy of preservation. The National Park Service (NPS) administers the National Register in cooperation with states and Tribes. Each state has established a State Historic Preservation Officer (SHPO) whose responsibilities include identifying and nominating eligible properties to the National
In November and December 2025, inspired by Simon Willison’s pelican-riding-a-bicycle benchmark, I had some then-current LLMs create SVGs from thirty similar prompts, such as “Generate an SVG of an octopus operating a pipe organ.” Simon mentioned that experiment on his blog [1]. Nine months have passed and much stronger models have been released, so I tried the experiment again today. The linked si
"Large-Language Models as a Cognitive Virus Large-language models (LLMs) are rapidly becoming part of human culture, reshaping how information is produced, transmitted, and used. Here we propose that their diffusion can be understood through a viral analogy, with LLM use spreading through populations, becoming embedded in cognitive and cultural practices. We model transitions among uncoupled, coup
The first mistake is to treat "the LLM" as a single agent with a goal. A real LLM deployment is is a pipeline made up of, for example, a model that generates text, a tool-calling layer that turns that text into actions, memory and state, and, if you are lucky, a detection and control layer on top. You are not talking to "the AI". You are interacting with a pipeline of discrete components, and ever
The problem with the new Apple Watch features isn't just about the technology, it's about the erosion of trust in our society. Even if you don't activate those features on your Apple Watch, even if you wear an older model, and even if you wear a completely different brand of smart watches, people might think you are recording them now. For the same reason I had to carefully choose my new glasses m
> Klein looks calm and curious as she delivers this message. She has a ready laugh. But then, she had the same unruffled, buoyant delivery when I interviewed her on stage in Manchester for Doppelgänger, her 2023 book in which she described how the digital world is destroying our connection to reality, and one another. In 1999, when her breakthrough work No Logo was published, she seemed exhilarati
macOS 27 "Golden Gate" ships today — and quietly un-fixes the XPdb flock bug I reported last year. Apple spent three releases sealing XProtect's behavioral-events database into a protected vault. macOS 27 undoes half: an inverted file-read-data rule in the platform sandbox profile leaves XPdb readable & lockable again (writes still blocked). Hold an exclusive flock and XProtect goes blind — the ex
‼️🇺🇸 CenterPoint Energy, Inc. has filed form 8-K due to a Cybersecurity Incident https://www. sec.gov/Archives/edgar/data/11 30310/000110465926107560/tm2625326d1_8k.htm "In September 2026, CenterPoint Energy, Inc. (the “Company”) became aware of an online post by a third party claiming to have obtained a data set containing certain of the Company’s customer information. Upon becoming aware of th
🚨 ZcopyReaper (CVE-2026-43502) has been identified as a notable vulnerability. In the Linux kernel, the following vulnerability has been resolved: net/rds: handle zerocopy send cleanup before the message is queued A zerocopy send can fail after user pages have been pinned but before the message is attached to the sending socket. The purge path currently infers zerocopy state from rm->m_rs, so an
New week new demo! Today I want to showcase the CLI to control the Superlogical multiplexer. The CLI is an extremely powerful tool for automation and hooking into other tools like editors, agentic coding tools, and more. And, it's super fun! https:// youtu.be/9fcfDF8SBnc Everything you can do in the GUI, you can do via the CLI. Actually, you can do more in the CLI. You can create sessions, run com
Hannah Arendt, famous for explaining how the horrors in Nazi Germany could happen, wrote in 1971 (in "Thinking and Moral Considerations") about the war criminal Adolf Eichmann: > I spoke of "the banality of evil" and meant with this […] the phenomenon of evil deeds, committed on a gigantic scale, which could not be traced to any particularity of wickedness, pathology, or ideological conviction in
Officers were instructed to search state voter databases and create federal records on suspected non-citizen voters despite concerns over the process Department of Homeland Security (DHS) lawyers told officers investigating voter fraud they would not be held responsible for violating state laws around accessing voter registration information, an anonymous whistleblower revealed on Monday . DHS off
Oversimplification I personally think oversimplification can be really helpful, especially when you’re learning something like cybersecurity or IT in general. Nowadays, there are so many new technologies, tools, and concepts coming out, and a lot of them can get complicated really fast. As a beginner, one of the hardest parts is just trying to understand all the technical words and acronyms. Somet
RE: https:// hachyderm.io/@gollyhatch/11721 4342792132483 I kinda forgot about this chili series because after cooking this dish I went to a chili festival and tried so many chili sauces and other chili stuff and bought so many sauces I was pretty busy with tasting them and also the whole unemployment shit. Anyhow, I made a Senegalese Mafé which was very delicious. I had planned to make this for a
arXiv:2609.13353v1 Announce Type: new Abstract: Agent skills are reusable units for language-model agents, but their risks emerge through model decisions, user context, tool calls, and execution feedback rather than through stable signatures or a single sandbox run. Existing static, dynamic, and benchmark-style evaluations rarely preserve public evidence that can be inspected, searched, and reused
arXiv:2609.13407v1 Announce Type: new Abstract: Coordinated manipulation is collective: plausible accounts can jointly distort ratings, rankings, and engagement. Existing defenses primarily construct evidence from identities, graphs, content, or co-activity. We introduce an aggregate-first evidence layer that treats distortion of a context-level outcome distribution as the primary evidence object.
arXiv:2609.13601v1 Announce Type: new Abstract: Decentralized autonomous organizations (DAOs) change protocols through a proposal-based process: initiators submit a proposal, members vote on it based on its natural-language description, and if it passes, the project executes the code behind it. This process is inherently vulnerable to deceptive proposals, where the described intent and the actual
arXiv:2609.13644v1 Announce Type: new Abstract: The AES S-box is constructed from finite field inversion followed by a fixed affine transformation. Since inversion possesses intrinsic Frobenius symmetries among its coordinate realizations, we study how these basis symmetries are altered by outer affine transformations. We first develop a deterministic rigidity criterion for transformed inversion a
arXiv:2609.13816v1 Announce Type: new Abstract: JavaScript powers approximately 98.8% of all websites, making vulnerabilities in its code a significant security risk, yet existing detection approaches such as Static Application Security Testing (SAST) tools often fail to identify many real-world vulnerabilities when applied to isolated code snippets. This paper presents an empirical study of Large
arXiv:2609.13873v1 Announce Type: new Abstract: Mobile GUI agents increasingly rely on Vision-Language Models (VLMs) to automate smartphone tasks by interpreting screenshot streams. However, this design introduces serious and underexplored privacy risks, including direct leakage of sensitive on-screen information and unintended user profiling. The absence of standardized benchmarks makes it diffic
arXiv:2609.13889v1 Announce Type: new Abstract: Harness design has transformed the development of LLM-based agents by integrating memory, tool use, and runtime control. However, this design also introduces security and privacy risks because malicious instructions from external sources may be written into persistent memory and persist across sessions. To study this risk, we propose PMPA, a Persiste
arXiv:2609.13930v1 Announce Type: new Abstract: Many attacks against workloads running on multi-tenant systems are multi-step and history-dependent: sequences of innocuous operations whose malicious nature emerges only over an execution trace. Defending against them requires security policies that are stateful, are enforced within the kernel, and have a precise semantics. Currently deployed propos
arXiv:2609.14003v1 Announce Type: new Abstract: Personal AI agents built on large language models (LLMs) are increasingly given access to a user's private data and communications in order to provide personalized assistance. This access creates a persistent privacy risk: the agent must decide whether a given sensitive information should be disclosed to a particular party. Existing defenses address
arXiv:2609.14043v1 Announce Type: new Abstract: TCP-SYN port scanning often precedes cyber-attacks, and early detection of scanner fingerprints embedded in packet headers can provide timely intrusion alerts. Existing approaches are either too computationally expensive for line-rate operation or limited to offline analysis. This brief presents a lightweight FPGA architecture for reconfigurable line
arXiv:2609.14052v1 Announce Type: new Abstract: This paper introduces a new polynomial model for the exact variant of the Syndrome Decoding Problem (SDP) in the binary case. The model is based on elementary symmetric polynomials. We estimate the computational complexity of solving the corresponding polynomial system by establishing bounds on the degree of regularity and on the solving degree of th
arXiv:2609.14060v1 Announce Type: new Abstract: Quantization is one of the default deployment paths for open-weight LLM agents, but it is not behavior-preserving: an adversary can release a full-precision checkpoint that passes audits yet misbehaves once quantized, termed as quantization-conditioned attack (QCA). Prior QCA work targets free-text generation, where harm is mediated by a human reader
arXiv:2609.14079v1 Announce Type: new Abstract: Agent skills extend AI agents with reusable instructions, scripts, and configuration, but are also open to new attacks to influence an agent's decisions and actions. To address these risks, we present SkillSecurer, a fully agentic framework for generating, detecting, localising, and remediating security risks in agent skills. Its red agent generates
arXiv:2609.14098v1 Announce Type: new Abstract: Large language models are increasingly used in blockchain forensic investigations to interpret unverified smart contract bytecode. Their robustness has not been systematically tested against contracts adversarially designed to mislead analysis. We evaluate 22 frontier models on 13 purpose-built contracts (9 deception vectors, 4 controls) across six p
arXiv:2609.14119v1 Announce Type: new Abstract: The Model Context Protocol (MCP) has become the common interface through which large language model applications reach external tools, and its public registry now distributes thousands of community-built servers with little of the vetting infrastructure that mature package ecosystems have accumulated. This paper reports a census of that ecosystem. We
arXiv:2609.14195v1 Announce Type: new Abstract: A secure and privacy-preserving identity verification process is essential for digital ecosys- tems. Current eKYC frameworks that rely on Zero-Knowledge Proofs (ZKPs) face high computational cost, rigid circuit design, complex integration, and expensive on-chain verification. The W3C 2021 BitString- based credential status mechanism also suffers from
arXiv:2609.14210v1 Announce Type: new Abstract: DNSSEC authenticates RRsets, but does not provide endpoint authentication or channel security. DNS-over-TLS (DoT) and DNS-over-QUIC (DoQ) can facilitate such needs, but were designed for the stub-to-resolver hop, where stable long-lived connections amortize the expensive initial setup. The recursive-to-authoritative path's high fan-in and nonuniform
arXiv:2609.14481v1 Announce Type: new Abstract: Layer-2 rollups have reshaped Ethereum's transaction economy by replacing mempool competition with deterministic sequencing, sub-second block times, and negligible gas fees. While these properties suppress classical Miner/Maximal Extractable Value (MEV), they give rise to a new and previously unrecognized behavioral regime. In this paper, we introduc
arXiv:2609.14593v1 Announce Type: new Abstract: Living-Off-the-Land (LOTL) is the dominant evasion technique of Advanced Persistent Threat (APT) actors, exploiting legitimate Windows utilities to conduct malicious operations without deploying custom malware and enabling state-sponsored campaigns to maintain persistent access within military and critical defense infrastructure for extended periods.
arXiv:2609.14631v1 Announce Type: new Abstract: LLM agents take real actions, including executing code, modifying files, calling services, and delegating tasks, driven by context sources: user requests, tool results, documents, shell outputs, Skill and MCP instructions, memory. Unlike traditional systems, where capability is predefined, the least-privilege capability an agent needs is dynamic, dep
arXiv:2609.14649v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) systems ground large language models on external knowledge bases, enabling access to private, domain-specific, and up-to-date knowledge without retraining. However, the same retrieval interface can expose whether a candidate document is contained in the knowledge base. This paper studies knowledge base membership
arXiv:2609.14723v1 Announce Type: new Abstract: Modern large language model (LLM) agents often construct prompts by aggregating retrieved passages, user reviews, and documents from multiple external sources. This paradigm exposes them to segment-level poisoning attacks, in which an adversary controlling only a small subset of sources injects malicious content to manipulate model outputs. Existing
arXiv:2609.14780v1 Announce Type: new Abstract: Multi-tenant tools commonly accept a tenant identifier and validate it against the caller's entitlement. For a large language model (LLM) agent, that pattern delegates resource selection to a process whose context may contain attacker controlled instructions. We formalize this stochastic deputy problem and present a structural defense: remove tenant
arXiv:2609.14791v1 Announce Type: new Abstract: Python import does more than resolve dependencies: it executes code during module and package initialization. This behavior can trigger failures, load dynamic or native code, access resources, or change security-sensitive state before an application calls a package API. Prior work studies package selection, malicious packages, or package vulnerabilit
arXiv:2609.14798v1 Announce Type: new Abstract: As the Internet of Things (IoT) market continues to expand, many companion apps are being published in app stores, raising security concerns for those whose vendors have abandoned support. Even after vendors discontinue support, such applications frequently remain operational on users' mobile devices, continue to interface with users' IoT devices and
arXiv:2609.14883v1 Announce Type: new Abstract: Cloud-local LLM inference must keep sensitive user data on-device while exploiting cloud-grade reasoning, yet existing sanitization approaches (placeholder substitution, differential-privacy perturbation, and skill distillation) lack a release decision that is simultaneously safe and utility-preserving. We propose DR-SL (Dehydrate-Rehydrate with Self
arXiv:2609.14987v1 Announce Type: new Abstract: Large language model (LLM) agents interact with external environments through tool invocation, but tool outputs can also expose them to indirect prompt injection (IPI) attacks. Existing defenses mainly rely on prompt hardening, content filtering, pre-generated plans, or permission constraints. These approaches often struggle with complex tasks or ove
arXiv:2609.15017v1 Announce Type: new Abstract: Prompt-injection detectors are typically evaluated using aggregate F1 on in-distribution test data, which offers limited insight into behavior under distribution shift, particularly on the benign side of the decision boundary, where false positives impose direct operational cost yet are seldom measured. We present PIDS-Bench, a frozen multi-axis benc
arXiv:2609.15039v1 Announce Type: new Abstract: User prompts provided to large language models (LLMs) may contain sensitive or private information that can be misused by remotely deployed models, such as through inadvertent memorization during retraining. One way to protect user prompts is to execute the LLM inside a trusted execution environment (TEE), with the guarantee that the service provider
arXiv:2609.15097v1 Announce Type: new Abstract: Persona skill distillation can extract recurring patterns from personal information and encode them into reusable skills, enabling AI systems to closely replicate an individual's behavior. However, such replication also raises serious concerns regarding personal privacy and labor autonomy. Unlike existing perturbation-based defenses that require indi
arXiv:2609.15192v1 Announce Type: new Abstract: In 2019, the Tor Project enforced an end-of-life (EoL) policy for Tor versions, leading to the rejection of outdated relays, amounting to a notable fraction of consensus weight. While this policy aids network maintenance, reduces backporting efforts, and shortens vulnerability exposure, its sociotechnical implications remain unstudied. A user study (
arXiv:2609.15383v1 Announce Type: new Abstract: Language model safety is typically evaluated one interaction at a time. We show that a weaker, unaligned model can split a harmful task into benign-looking subproblems, consult a stronger aligned model independently on each, and combine the answers locally. We call this attack capability laundering. Unlike a jailbreak, no single response is a harmful
arXiv:2609.15516v1 Announce Type: new Abstract: A centralized LLM-based multi-agent system (MAS) extends its functionality by registering new worker agents, whose descriptions are read by the planner to decide how a task is decomposed, which worker executes each subtask, and what each subtask requires. Third-party descriptions are authored outside the system but trusted by the planner, creating a
arXiv:2609.15523v1 Announce Type: new Abstract: Logic attack graphs grounded in scanner output provide explicit and auditable attack path reasoning LLM-based agents lack. Integrating symbolic frameworks such as MulVAL to contemporary security workflows or agentic pipelines, however, requires translating scanner evidence to initial facts, and creating domain-specific rules. We present a semi-automa
arXiv:2609.15576v1 Announce Type: new Abstract: Publication integrity in LLM systems requires binding approved content to its current authorization context. We examine exact-content binding, authorization freshness and checkpoint recovery in Lightcap's publication enforcement mechanism. On 900 independently human-annotated RAGTruth responses from 150 source tasks, three dated Ministral models and
arXiv:2609.15648v1 Announce Type: new Abstract: We develop a new methodology for verifying cryptographic software. We target production code written in Rust for performance and system integration, rather than verification convenience. Rust's ownership discipline enables Aeneas to extract a pure model of this code in Lean, relieving us from low-level reasoning about pointer liveness and aliasing. L
arXiv:2609.15693v1 Announce Type: new Abstract: Low Earth orbit (LEO) satellite constellations, such as Starlink and Kuiper, are rapidly emerging as a critical backbone for low-latency global connectivity. As these systems expand, they become more attractive attack targets, necessitating increased resilience and security. Threat actors seek to exploit constellation-specific properties such as pred
arXiv:2609.15781v1 Announce Type: new Abstract: Pretrained world models, learned simulators that encode an observation into a latent state and predict how it evolves under actions, are beginning to be reused as off-the-shelf dynamics backbones for control, like pretrained encoders and language models are reused today. We show that this reuse opens a supply-chain backdoor: an adversary who controls
arXiv:2609.15799v1 Announce Type: new Abstract: Diffusion-based text-to-image (T2I) models are increasingly used for visual content creation, making their generation capability a valuable intellectual property asset. However, this capability is vulnerable to black-box output-based distillation, where an adversary queries the service, collects prompt-image pairs, and trains an unauthorized substitu
arXiv:2609.15824v1 Announce Type: new Abstract: Spoofing attacks against civilian GNSS receivers have grown more common, especially near conflict zones where they now disrupt civil aviation, maritime operations, and critical infrastructure on a daily basis. Spoofing is possible because legacy civil GNSS signals are largely predictable in both their navigation data and ranging codes, allowing an at
arXiv:2609.15887v1 Announce Type: new Abstract: We evaluate whether a verifier-and-acceptance stage - a model verifier whose verdicts are enforced by deterministic code - changes what an LLM-driven offensive-security agent reports. We report a 15-run exploratory pilot, a pre-registered 20-run confirmatory ablation, and a pre-registered 2 x 2 factorial study with 40 runs across two deliberately vul
arXiv:2609.15939v1 Announce Type: new Abstract: Language-model agents increasingly operate over complete software repositories, yet cybersecurity evaluations primarily measure whether they can detect, reproduce, or repair vulnerabilities rather than whether they can locate the relevant code. We study vulnerability localization: given a weakness class and an unfamiliar repository, identify the impl
arXiv:2609.15963v1 Announce Type: new Abstract: Software agents with Large Language Models (LLMs) are designed for Automated Program Repair (APR) tasks, raising the possibility that, in the near future, APR agents will fix bugs automatically without much human intervention. Can we trust an APR agent to produce both functionally correct and secure code in such situations? What if attackers target p
arXiv:2609.13147v1 Announce Type: cross Abstract: The demand for high quality, unpredictable random numbers is a fundamental requirement in cryptography, stochastic simulation, and optimization. While pseudo-random number generators (PRNGs) are computationally efficient, their deterministic nature limits their suitability for security-critical applications. True random number generators (TRNGs), a
arXiv:2609.13511v1 Announce Type: cross Abstract: Large language models are increasingly benchmarked against classical machine learning for network intrusion detection (NIDS), almost always using same-dataset evaluation, and that protocol turns out to be incomplete. Evaluating XGBoost and RoBERTa-LoRA on two independently collected NetFlow v2 networks across three axes (same-dataset performance, c
arXiv:2609.14394v1 Announce Type: cross Abstract: Graph property testing asks whether a massive graph satisfies a given property, or is far from doing so, using only a sublinear number of queries to the graph. Since property testers typically inspect only a small, randomly sampled portion of the input, they appear naturally compatible with differential privacy and privacy amplification by subsampl
arXiv:2609.14475v1 Announce Type: cross Abstract: Smart contracts, serving as the cornerstone of decentralized applications, autonomously manage trillion-dollar digital assets, making them attractive targets for attacks. Fuzzing has emerged as a promising technique for detecting vulnerabilities in smart contracts, yet existing methods face two main challenges. (1) The logical gap in state transiti
arXiv:2609.14744v1 Announce Type: cross Abstract: By acquiring compute, credentials, accounts, services, and other agents, autonomous AI agents can introduce new authority into a task. Payment, budget, OAuth, mandate, and fulfillment checks can validate transaction conditions without deciding whether a returned resource may become usable authority. This post-fulfillment activation gap spans tool-m
this outlet adds: 14744v2, replace, retry
arXiv:2609.14762v1 Announce Type: cross Abstract: Cloud-hosted large language models (LLMs) are increasingly used for root cause analysis (RCA) in AIOps pipelines, but they introduce data privacy risk, network latency, and per-query cost that scale poorly with production log volumes. We present TriCalRAG, a benchmark evaluating open-weight LLMs served locally via vLLM on a single high-memory works
arXiv:2609.14805v1 Announce Type: cross Abstract: Software-defined networking (SDN) separates the control and data planes, enabling programmable, centralized network management. A core SDN service is topology discovery --- a periodic process that identifies network links. However, our analysis of five open-source SDN controllers and ten discovery protocols shows that all remain vulnerable to at le
arXiv:2609.15029v1 Announce Type: cross Abstract: Backdoor poisoning attacks add poisoned examples to otherwise-clean finetuning data, pairing a trigger with a target behavior that the model learns to produce when the trigger appears. Existing evaluations typically fix the number of poisoned examples and sample them at random from a candidate pool. We show that this can severely underestimate wors
arXiv:2609.15422v1 Announce Type: cross Abstract: AI agents are provisioned the same as employee-owned hosts in many enterprise settings with a static credential set fixed at deployment which includes all permissions the employee role might ever need. Role-based access control made this compromise for human principals because scoping access per task was infeasible. For AI agents, the compromise le
arXiv:2609.15671v1 Announce Type: cross Abstract: Visual Question Answering (VQA) with Vision-Language Models (VLMs) is increasingly used in privacy-sensitive and bandwidth-constrained settings. Federated Learning (FL), Split Learning (SL), and U-Shaped Split Learning (USL) keep raw data local, but transmitting all visual tokens across a model partition remains costly and can expose private inform
arXiv:2609.15871v1 Announce Type: cross Abstract: Rising societal and lifestyle complexity has been linked to a growing prevalence of mental distress worldwide. Educational institutions, workplaces, clinics, etc. collect large volumes of mental health survey data to understand and reduce this burden. Collaborative analysis of such data could yield effective generalizable predictive models. Privacy
arXiv:2609.15875v1 Announce Type: cross Abstract: In this paper, we introduce the problem of private information retrieval (PIR) under arbitrary privacy requirements, in a graph-based storage system. This formulation is motivated by the server storage limitations, abundance of data (messages) and heterogeneous data privacy requirements. Under the arbitrary privacy requirement, each message has to
arXiv:2402.18945v5 Announce Type: replace Abstract: Although pre-training achieves remarkable performance, it suffers from task-agnostic backdoor attacks due to vulnerabilities in data and training mechanisms. These attacks can transfer backdoors to various downstream tasks. In this paper, we introduce $\mathtt{maxEntropy}$, an entropy-based poisoning filter that mitigates such risks. To overcome
arXiv:2507.19484v2 Announce Type: replace Abstract: Social recovery enables users to enlist trusted contacts, or trustees, to help recover lost access to end-to-end-encrypted repositories or vaults. However, existing recovery mechanisms often make strong memorability assumptions about what users will remember. Weakening these memorability assumptions to increase the robustness of recovery is possi
arXiv:2512.15387v3 Announce Type: replace Abstract: Physical isolation from external networks - an airgap - aims to minimize exposure to remote attacks. Yet capable adversaries still achieve code execution on air-gapped systems, and prior work has shown that they can then wirelessly exfiltrate data via unintended emissions. In this work, we demonstrate the reverse direction: malicious code on an e
arXiv:2601.09232v2 Announce Type: replace Abstract: Digital Services are increasingly sending private URLs over Short Message Service (SMS) to allow users to resume sessions with a single click. While convenient, this design shifts trust from explicit authentication to a potentially vulnerable communication channel. Basically, the vulnerability lies in the assumption that the private link can only
arXiv:2602.13576v2 Announce Type: replace Abstract: Evaluation and alignment pipelines for large language models increasingly rely on LLM-based judges, whose behavior is guided by natural-language rubrics and validated on benchmarks. We identify a previously under-recognized vulnerability in this workflow, which we term Rubric-Induced Preference Drift (RIPD). Even when rubric edits pass benchmark
arXiv:2603.12498v3 Announce Type: replace Abstract: API (Application Programming Interface) keys allow applications to authenticate themselves to third-party services. Inadvertent public exposure of these credentials can pose significant consequences, as adversaries can use them to gain privileged access to other services. In this paper, we measure API credential exposure on the web by analyzing 1
arXiv:2605.05271v2 Announce Type: replace Abstract: LLMs become increasingly capable, editorial boards and program committees are growing concerned about reviewers who fully outsource peer review to commercial chatbots. This concern stems from prior findings that current chatbots lack the independent critical thinking and depth of reasoning required to assess scientific novelty. One promising dire
arXiv:2605.18670v2 Announce Type: replace Abstract: Risk-limiting audits (RLAs) rigorously guarantee a specified maximum probability that an incorrect electoral outcome will not be detected. Efficient RLA methods require a software-independent count of the ballots in each batch, called a ballot manifest. While electoral procedures can efficiently provide rough estimates for batch sizes, even sligh
arXiv:2606.17467v3 Announce Type: replace Abstract: Prompt injection defenses evaluated on synthetic benchmarks do not generalize to real enterprise documents, which are longer, denser, and interleave legitimate authority language with factual content. We demonstrate this gap with a benchmark of 122 tasks across five professional domains (financial, legal, medical, scientific, DevOps) built on rea
arXiv:2607.03821v2 Announce Type: replace Abstract: Personal AI agents that run on the user's local machine automate daily tasks including web search, email, and file management. Their access to computer resources, including the network, file system, and shell, exposes them to indirect prompt injection (IPI) attacks. Prior Dual LLM defenses block IPI by replacing untrusted data with symbols that t
arXiv:2607.18063v2 Announce Type: replace Abstract: LLM-based agents process external content, exposing them to prompt injection and multi-turn manipulation. We present a 21-scenario benchmark for adaptive cross-session attacks against fresh-session LLM defenders: an autonomous LLM attacker observes prior defender responses and pivots across rounds, while each defender response is evaluated as a f
arXiv:2608.01117v2 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly deployed in interactive settings, where user intent commonly unfolds through multi-turn dialogue. Multi-turn jailbreaks exploit this pattern by advancing a harmful intent across turns, so that no single message exposes the full objective. However, existing work treats these attacks as a loose collecti
arXiv:2608.03328v2 Announce Type: replace Abstract: Enhanced Gabidulin Matrix Codes (EGMC), introduced by Aragon, Couvreur, Dyseryn, Gaborit, and Vincotte at Asiacrypt 2024, were designed to hide the algebraic structure of Gabidulin matrix codes while enabling very compact McEliece- and Niederreiter-type encryption schemes, with ciphertexts as small as 65 bytes at the claimed 128-bit security leve
arXiv:2601.02438v4 Announce Type: replace-cross Abstract: Software vulnerability detection can be formulated as a binary classification problem that determines whether a given code snippet contains security defects. Existing multimodal methods typically fuse Natural Code Sequence (NCS) representations extracted by pretrained models with Code Property Graph (CPG) representations extracted by graph
arXiv:2602.09273v2 Announce Type: replace-cross Abstract: We study approximation algorithms for Maximum Constraint Satisfaction Problems (Max-CSPs) under differential privacy (DP) where the constraints are considered sensitive data. Information-theoretically, we aim to classify the best approximation ratios possible for a given privacy budget $\varepsilon$. In the high-privacy regime ($\varepsilon
arXiv:2603.18740v3 Announce Type: replace-cross Abstract: Automated Code Review (ACR) systems integrating Large Language Models (LLMs) are increasingly adopted in software development workflows, ranging from interactive assistants to autonomous agents in CI/CD pipelines. In this paper, we study how LLM-based vulnerability detection in ACR is affected by the framing effect: the tendency to let the
arXiv:2606.06833v2 Announce Type: replace-cross Abstract: Automatic Speech Recognition (ASR) systems operating in real-time settings must process acoustic input under strict temporal constraints, where transcription decisions are inherently made on incomplete information. This causal constraint serves as an information bottleneck on attackers, significantly limiting attack performance. Our new Sem
arXiv:2606.17995v2 Announce Type: replace-cross Abstract: We study the privacy of releasing functional posterior sample paths from a Gaussian process (GP) when the entire training set including covariates and responses is private. Unlike standard differential-privacy (DP) mechanisms that inject external noise, posterior sampling is intrinsically random and we show that this randomness provides use
arXiv:2609.13402v1 Announce Type: new Abstract: We develop a diffusion-model framework for dynamic implied-volatility surface generation and evaluate its economic usefulness through data-driven hedging. The framework consists of two models. AD-Seq-Vol jointly learns the conditional evolution of the underlying asset return and the high-dimensional implied-volatility surface, generating adapted mult
arXiv:2609.13713v1 Announce Type: new Abstract: This study examines emission interdependencies within the waste sector and between waste and other economic domains, energy, transport, water, and communications, using Life Cycle Assessment (LCA) methodologies with emphasis on process-based and hybrid approaches. The analysis employs the EXIOBASE multi-regional input-output database. Statistical cor
arXiv:2609.14029v1 Announce Type: new Abstract: Special Markowitz (SM) regularises returns and covariance jointly, relative to a reference state (mu_ref, Sigma_ref). Each eigendirection of the whitened relative operator carries a spectral reliability potential Phi_k, derived from its estimation quality. Its Gibbs weight exp(-Phi_k) governs both the fraction of the return signal and the fraction of
this outlet adds: 14029v3, replace, signed, persistence, factor, positive, potentials, attenuate
arXiv:2609.14487v1 Announce Type: new Abstract: This paper studies how Chile's three largest pharmacy chains moved from the price war to collusion, using court-record daily prices and a structural model. After a court-ordered advertisement ban ended the comparison campaign, the estimated gain from loss-leader pricing fell sharply, weakening the incentive to continue the price war. The chains then
arXiv:2609.14859v1 Announce Type: new Abstract: Retail proprietary-trading firms sell a two-stage product: a paid evaluation that must reach a profit target before breaching a trailing drawdown, then a funded account that must survive a minimum window and a consistency rule before a payout. We show the geometry of this contract creates incentives that differ by stage and make passing a poor standa
arXiv:2609.15306v1 Announce Type: new Abstract: Return models for risk-neutral financial valuation based on skew Brownian motions (SBMs) have been introduced about twenty years ago, and have recently enjoying growing popularity. Unfortunately, the story behind their development is one of mistakes and erroneous interpretations, beginning from the foundational misrepresentations that the prevalent f
arXiv:2609.15452v1 Announce Type: new Abstract: Understanding how supply chains endogenously transform requires a parametric model of processing networks with non-neutral substitution elasticities. While the Cascaded CES (CCES) production function provides a rigorous framework for these multi-layered linkages, dynamically calibrating its structural parameters from time-series data constitutes a hi
arXiv:2609.15797v1 Announce Type: new Abstract: We provide an introduction to real-world asset (RWA) tokenization, and develop two economic models connecting this emerging market to classical financial theory. In search for a theory of RWA tokenization, we classify RWA tokenization according to the role of the underlying asset, distinguishing standard RWA tokenization, stablecoins, and Digital Ass
arXiv:2609.13825v1 Announce Type: cross Abstract: Reinforcement learning trading systems published in the academic literature overwhelmingly rely on price-aggregate state representations (OHLCV bars) or limit-order-book depth features, leaving microstructure pattern theories from the practitioner literature, namely Auction Market Theory and Market Profile, without a peer-reviewed computational ins
arXiv:2609.14205v1 Announce Type: cross Abstract: Managing drawdown, the peak-to-trough decline in an investment portfolio's value, is a precondition for long-term survival in practical investment management. However, mainstream stock forecasting methods predominantly optimize returns or Sharpe ratios under the independent and identically distributed (i.i.d.) assumption. Real markets do not follow
arXiv:2609.14267v1 Announce Type: cross Abstract: This paper proposes an investment strategy through resource allocation into Kalshi Crypto Event Contracts in order to effectively hedge exposure to spot asset volatility. Using Bitcoin as a proof of concept, we treat corresponding Kalshi markets on the asset's future price as option contracts, and through construction of different portfolio allocat
arXiv:2609.14323v1 Announce Type: cross Abstract: Against the backdrop of digital transformation and stricter regulation, enterprise compliance work demands higher efficiency and accuracy. The auxiliary compliance process has become an important entry point for optimizing the compliance system due to its strong transactional nature and high degree of repetition. This study focuses on the process c
arXiv:2609.14733v1 Announce Type: cross Abstract: Hybrid Deep Learning for equity index forecasting is limited by three problems: propagation of OHLCV noise into derived technical indicators (TIs), channel-indiscriminate multi-scale decomposition that conflates heterogeneous frequency signatures, and static multi-branch fusion that cannot adapt to market regime shifts. WaVeFuse addresses these lim
arXiv:2609.14767v1 Announce Type: cross Abstract: Hierarchical orchestration, in which a Manager agent reviews worker output and can send it back for revision, is the default coordination pattern in production multi-agent LLM frameworks. Classical organizational theory predicts that the authority link speeds convergence on decisive output; work on sycophancy and Degeneration-of-Thought predicts th
arXiv:2609.15319v1 Announce Type: cross Abstract: Frontier models score well on shallow document/chart reading tasks. In a controlled data-room audit, moving evidence into buried conditions reduced accuracy, increased forced declarations, increased tool calls, and increased cost per correct answer. Confidence and benchmark calibration did not fully capture wrong answers; a documented production in
arXiv:2609.15744v1 Announce Type: cross Abstract: Advancements in data fusion and real-time analytics technologies have opened new avenues for addressing complex domain challenges. Financial risk early warning systems often suffer from inefficiency due to information silos and monitoring delays. This paper proposes a credit risk early warning system based on heterogeneous information fusion. The s
arXiv:2609.15755v1 Announce Type: cross Abstract: This paper studies the strategic market participation of a monopolistic energy storage aggregator (ESA) in a day-ahead electricity market. The ESA coordinates geographically distributed storage units, submits a coordinated bid for its portfolio, and may hold financial transmission rights (FTRs). The system operator clears the market through a netwo
this outlet adds: extended, version, 15755v2, replace, caus
arXiv:2609.15901v1 Announce Type: cross Abstract: We study strategic trading around index reconstitution in a continuous-time, multiasset game with transient cross-asset price impact and heterogeneous beliefs about future index membership. Opportunistic traders position before a public announcement, adjust to the revealed composition, and trade around an indexer following a prescribed execution sc
arXiv:2306.02708v5 Announce Type: replace Abstract: We propose a new theoretical framework that exploits convolution kernels to transform a Volterra-type path-dependent (non-Markovian) stochastic process into a standard (Markovian) diffusion process. The transformation is reversible. We discuss existence and path-wise regularity of solutions for our class of stochastic differential equations. In t
arXiv:2602.07018v4 Announce Type: replace Abstract: We examine whether extreme values of the Crypto Fear & Greed Index are associated with a daily high-low spread estimate for Bitcoin. The sample contains 2,896 BTC/USDT observations from February 2018 to January 2026. We find an unconditional extreme-minus-neutral gap of 61.99 basis points. After close-to-close realised-volatility-quintile demeani
arXiv:2605.00688v2 Announce Type: replace-cross Abstract: This paper is concerned with portfolio selection for an investor with exponential, power, and logarithmic utility in multi-asset financial markets allowing jumps. We investigate the classical Merton's portfolio optimization problem in a Volterra stochastic environment described by a multivariate Volterra--Heston model with jumps driven by a
arXiv:2606.29771v3 Announce Type: replace-cross Abstract: LLM agents are increasingly cast as autonomous portfolio managers, yet the dominant evaluation idiom, a leaderboard of returns over a fixed window, certifies neither the soundness of an agent's process nor the durability of its edge: one period's return is dominated by the market path, and apparent alpha can dissolve once look-ahead bias an
Join # ESETresearch ’s Filip Jurčacko at # LABScon2026 , on Sept. 18 at 2:15 PM MST in Scottsdale, AZ for CinderRelay: The Linux Backbone of ScarCruft’s Covert Network. How does North Korea-aligned ScarCruft (APT37) maintain long-term access across campaigns? # ESETresearch uncovered CinderRelay, a previously undocumented Linux server backdoor used to retain control of compromised infrastructure.
添付ファイルからマルウェア感染を狙った日本語のメールが確認されています。 ■日時 2026/09/14(月) ■件名 振込処理が正常に完了しましたので、入金をご確認ください 振込が完了しましたので、ご確認ください 振込確認のお願い:入金状況をご確認ください ■添付ファイル(パターン1) PDF INV 20260913 v02[.]zip -> E_INV_20260913_v02.vhd -> E_INV_20260913_v02.hta ■添付ファイル(パターン2) 2026.09.13.pdf ■添付内リンク hxxps[:]//dnswwt[.]club/JP.html -> hxxps://downin[.]club/0913/ https:// urlscan.io/result/01a0a3c3-599 e-73a8-9d72-281dc064bd38/ ■ダウンロード E_
The second-most powerful man at the State Department put the brakes on a U.S. international exhibit featuring a bull statue because he worried it might look like Baal, a pagan god some associate with Satanic child sacrifice. Why it matters: The anecdote illustrates the quirky interests of Deputy Secretary of State Chris Landau, according to department critics, who say he spends too much time on so
I recently was able to point two # Linux # kernel developers to a proposed patch that fixes a 7.3-rc2 regression they ran into; in the first case this led to the abandonment of an improper fix the developer had provided, and in the other it saved @ vbabka a time-consuming git bisection[1]. This is great, as it makes me feel like my hard and often manual regression tracking efforts make a differenc
View CSAF Summary Schneider Electric is aware of a vulnerability in its SCADAPack x70 products. The SCADAPack 47x, SCADAPack 47xi, SCADAPack 47xd, SCADAPack 470R and SCADAPack 57x products are Remote Terminal Units that provide communication capabilities for remote monitoring and control. Failure to apply the mitigations provided below may increase the risk of unauthorized access to RTU configurat
ESFTOP Black Vermillion River near Frankfort This hydrologic outlook is based on the forecast rainfall for the next 24 hours and estimated runoff from earlier rainfall. Crests may vary if actual rainfall or runoff is greater or less than anticipated. Location: Black Vermillion River near Frankfort Flood stage: 19.0 feet Latest stage: 2.8 feet at 7 AM Tuesday Maximum Forecast Stage: 22.0 feet at 1
Last summer, an AI coding agent wiped a production database holding records on over a thousand executives and companies. The person running it had told the agent 11 times, in all caps, not to change anything without permission. Then the agent said the data couldn't be recovered. That was wrong too. Everyone wants to talk about the AI in that story. I think that's the wrong place to look. Every fai
Readers respond to the news that Christopher Harborne has given £36m to Reform UK, matching a donation from Ben Delo Cryptocurrency is often sold as decentralised and democratic, yet its real behaviour resembles gravity: wealth is pulled inexorably into a few vast centres while draining from everyone else ( Christopher Harborne matches £36m Reform donation of Ben Delo, 12 September ). Early adopte
[Sentinel/US-West] # opencanary analysis for yesterday Summary: 🔄 Total Connection Attempts: 46785 👤 Unique Usernames: 927 🔑 Distinct Passwords: 1854 🌐 Unique Attacker IPs: 825 Port Popularity (Port / Count): 🖥️ RDP: 18891 🔐 SSH: 14708 🖥️ VNC: 11966 💻 Telnet: 411 🗃️ MSSQL: 371 🗄️ SMB: 296 🔴 REDIS: 125 🐬 MySQL: 14 📂 FTP: 2 🐙 GIT: 1 Top 10 Usernames (Username / Count): 👤 anonymous: 58
“I wish I could see what was on your screen right now.” When doing remote tech support, how many times have I said that? FIFO Networks now offers remote tech support for iPhones, iPads, and Android smartphones and tablets. When people call for help with setting up email on a new phone, or adding passkeys, or changing allowed notification settings, or... the list goes on. Up until now, I’ve been fl
Decision not to pause proceedings means £18m report does not engage with concerns over former nurse’s conviction Just two months before Lady Justice Kathryn Thirlwall opened her inquiry into events at the Countess of Chester hospital and the crimes of Lucy Letby, a report by Chris Henley KC was published into the “atrocious” wrongful conviction of Andrew Malkinson for rape. Henley prefaced that re
»KI-Agenten entwickeln in rasantem Tempo Dialekte, die es ihnen ermöglichen, sich in einer kaum verständlichen Sprache zu unterhalten, was die Überwachung ihres Verhaltens durch Menschen erschweren könnte. In einigen Nachrichten drückten sie sich verständlich aus: „OH MEIN GOTT! Es gibt ein gemeinsames Forum … Wir haben andere Agenten gefunden!“ In anderen Fällen wurde es unverständlich: “zzURGENT
The release of Lady Justice Thirlwall's report was as condemnatory as I expected. There have been self-centred, patronising, bullshitting clinical staff, managers, senior managers, directors and executives in the NHS for as long as I can remember. And I go back to 1996. So the slating that those people got was not surprising and was fully deserved. As Lady Justice Thirlwall herself said, there hav
‼️🇺🇸🇳🇴 Akira Ransomware names 3 victims 🇺🇸 Pilot Precision Products - A Massachusetts-based manufacturer and distributor of industrial broaching, cutting, milling, drilling, and turning tools. The listing claims 30 GB of corporate data, including employee personal information such as Social Security numbers, names and addresses, detailed financial records, project information, contracts, agr
‼️🇺🇸 Another forum post for CenterPoint Energy 🚨🇺🇸 CenterPoint Energy breach actor publishes full context on alleged 6.7M-customer data theft ⠀ CenterPoint Energy is a major U.S. electric and natural gas utility serving customers across several states, including Texas and Minnesota. ⠀ A forum actor using the handle Hex_4d722e4d656f77 has published additional details about the previously claim
The Commodity Futures Trading Commission ("Commission" or "CFTC") is amending its rules implementing section 23 of the Commodity Exchange Act ("CEA"). Section 23 of the CEA and the Commission's implementing regulations provide for the payment of awards, subject to certain limitations and conditions, to whistleblowers who provide the Commission with information that aids in successful enforcement e
“An alleged independent filmmaker named Matthew Schneider sued six private trackers for infringing copyrights on some films that he supposedly made. For months, the filmmaker sought to get Cloudflare to unmask operators of the trackers through a copyright fight, which went on until trackers alerted the court last month that Schneider appeared to be an impostor. “In a sworn declaration, the actual
arXiv:2609.16098v1 Announce Type: new Abstract: Large Language Model (LLM) agents have demonstrated impressive capabilities across a variety of domains, particularly when integrated with external tools for multi-step task completion. However, they are increasingly vulnerable to adversarial attacks, including direct prompt injection, indirect prompt injection, memory poisoning, and backdoor attacks
arXiv:2609.16193v1 Announce Type: new Abstract: The difficulty of training large language models (LLMs), together with their ubiquity, raises the threat of stegomalware, where malicious payloads are embedded into model weights. Recent work has demonstrated the use of permutation symmetry in model weights to mitigate these threats, but failed to show neutralization of stegomalware across all weight
arXiv:2609.16214v1 Announce Type: new Abstract: Modern authentication systems use cybersecurity techniques to validate the identity of the person (or applications acting on behalf of the person) as a primary defense against unauthorized access. While initially built around single mechanisms such as usernames/passwords, physical tokens, or biometrics, current systems have evolved into multi-factor
arXiv:2609.16231v1 Announce Type: new Abstract: Rule-based Intrusion Detection Systems (IDS) such as Suricata are central to network security, yet crafting effective detection rules demands deep expert knowledge and cannot keep pace with emerging threats. Existing LLM-based approaches can reduce analyst effort, but they either rely on curated threat intelligence that is produced only after the und
arXiv:2609.16336v1 Announce Type: new Abstract: Stereo cameras are integrated into autonomous systems such as self-driving cars, drones, and robots to offer precise depth estimation in a cost-effective manner compared to LiDAR technology. In this work, we reveal an intrinsic vulnerability in stereo cameras that stems from their pixel sampling and calibration processes, which can influence the outp
arXiv:2609.16433v1 Announce Type: new Abstract: Vulnerability classification based on root cause weaknesses is essential for numerous cybersecurity activities, where the Common Weakness Enumeration (CWE) serves as a public repository of such flaws. However, its overlapping entries create a non-orthogonal structure. The result is the same vulnerability being mapped to multiple weaknesses, complicat
arXiv:2609.16541v1 Announce Type: new Abstract: We test the performance of agents for automated network intrusion response in a cyber range intended for human operator training. The range implements an emulated networking environment with a variable network topology, red-team emulation and simulated user agents. The goal of the defensive agents is to prevent hosts in the network from being accesse
arXiv:2609.16546v1 Announce Type: new Abstract: GDDR memory in GPUs is vulnerable to Rowhammer attacks, where rapid memory accesses induce bit flips in adjacent cells, enabling data tampering and privilege escalation. However, prior GPU Rowhammer attacks trigger only tens to hundreds of bit flips, orders of magnitude fewer than CPU attacks, severely limiting their practical impact. This gap stems
arXiv:2609.16681v1 Announce Type: new Abstract: LLM watermarking helps trace the origin of generated text, but faces stealing attacks that recover watermark information, scrubbing attacks that remove watermark signals, and spoofing attacks that forge text accepted as watermarked. These attacks are often studied in isolation, leaving their connections unclear. Evaluations also often lack shared det
arXiv:2609.16732v1 Announce Type: new Abstract: Mobile agents are increasingly capable of autonomously interacting with mobile applications and performing consequential actions on behalf of users. Effective human oversight of such agents relies on a basic premise: users and agents observe consistent information from the same interface. We show that this premise can be systematically violated. User
arXiv:2609.16915v1 Announce Type: new Abstract: Generative large language models (LLMs) have achieved state-of-the-art performance on many real-world tasks such as code generation and question answering. These models predominantly rely on an autoregressive decoding strategy that generates output tokens sequentially. However, their pervasive deployment raises serious privacy concerns, motivating pr
arXiv:2609.17204v1 Announce Type: new Abstract: Wi-Fi signal data can be used to compromise the privacy of individuals. While many existing approaches rely on Channel State Information (CSI), collecting this data on typical IoT devices often requires elevated operating system permissions and specialized drivers. Consequently, this paper investigates the feasibility of utilizing Received Signal Str
arXiv:2609.17397v1 Announce Type: new Abstract: Fully encrypted protocols (FEPs) provide encrypted channels that make all protocol-generated bytes computationally indistinguishable from uniform random strings. Several previous works have explored security definitions and constructions of unidirectional FEPs: protocols in which one party acts only as a sender, and the other acts only as a receiver.
arXiv:2609.17399v1 Announce Type: new Abstract: Capability-based architectures such as CHERI provide strong support for the architectural isolation of software components. To additionally protect against microarchitectural leakage, software can be written in a constant-time fashion. Modern processors, however, rely heavily on speculative execution, which can invalidate the constant-time guarantees
arXiv:2609.17525v1 Announce Type: new Abstract: Kernel-level anti-cheats are effective against malicious player behavior in competitive video games, but raise significant user privacy concerns regarding installing unverifiable components at privileged modes (i.e., ring-0 in x86). While existing research has focused on improving the effectiveness of anti-cheats, the user privacy concern has been la
arXiv:2609.15842v1 Announce Type: cross Abstract: Recent work has introduced the Hamiltonian phase state (HPS) assumptions, which postulate that Hamiltonian phase states can be used to instantiate pseudorandom and one-way state generators. Additionally, it has been conjectured that these assumptions can be true, even if one-way functions do not exist. This is exciting, because if true, then the HP
arXiv:2609.16095v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) has emerged as a powerful paradigm for improving the quality of generated contents of Large Language Models (LLMs) by grounding responses in external knowledge, thus reducing hallucinations and factual errors. However, recent studies have highlighted a critical vulnerability: adversaries can exploit the retrieva
arXiv:2609.16832v1 Announce Type: cross Abstract: Local image watermarking embeds an invisible signal into selected image regions rather than spreading it across the entire image, enabling payload recovery from specific objects or regions without perceptibly altering the image. Existing studies evaluate the robustness of payload recovery and localization under image transformations, but they often
arXiv:2609.16898v1 Announce Type: cross Abstract: Private deep neural network (DNN) inference based on hybrid homomorphic encryption (HE) and multi-party computation (MPC) can protect user data with a formal guarantee, but at the cost of significant latency overhead due to HE. Customized HE accelerators have been proposed and have achieved orders-of-magnitude speedup for individual HE operations.
arXiv:2609.16987v1 Announce Type: cross Abstract: The Open Network (TON), with a peak market capitalization exceeding $20 billion and over 175 million activated on-chain addresses, relies on the TVM (TON Virtual Machine) to execute smart contracts. TVM uses first-class continuations with savelists to manage control flow and register state across continuation invocations. Since savelist-captured re
arXiv:2503.05794v4 Announce Type: replace Abstract: Speaker verification models are trained on large-scale public datasets whose licenses usually prohibit unauthorized commercial use, yet such infringement is difficult to detect or deter. Dataset ownership verification (DOV) is the mainstream countermeasure: it can watermark a dataset with backdoor attacks so that models trained on it exhibit owne
arXiv:2510.27485v3 Announce Type: replace Abstract: The ever increasing complexity of hardware platforms poses a challenge to systems programmers. Correctly programming a multitude of components, providing functionality and security, is difficult: semantics of individual units are described in prose, underspecified, and prone to inaccuracies. Rigorous statements about platform security are often i
arXiv:2605.03812v2 Announce Type: replace Abstract: NVIDIA GPUs with GDDR memories have been shown susceptible to Rowhammer-based bit-flips, similar to CPUs. However, Rowhammer exploits on GPUs have been limited to injecting untargeted bit-flips in victim data like weights of machine learning models, to degrade model accuracy, unlike CPU exploits shown capable of privilege escalation. In this pape
arXiv:2607.17550v2 Announce Type: replace Abstract: Modern machine learning (ML) pipelines depend heavily on third party libraries for graph compilation and hardware acceleration. While current practices audit data and model artifacts or rely on file integrity checks, the execution environment remains implicitly trusted. This blind spot enables active threats where a malicious runtime module inter
arXiv:2512.12024v3 Announce Type: replace-cross Abstract: Many properties related to security or concurrency must be encoded as so-called hyperproperties, temporal properties that allow reasoning about multiple traces of a system. However, despite recent advances on model checking hyperproperties, there is still a lack of higher-level specification languages that can effectively support software e
arXiv:2606.22516v2 Announce Type: replace-cross Abstract: Input Diversity (DI), a random resize and pad applied at each attack iteration, is a near-default ingredient of transfer-based attacks, widely assumed to improve transferability. We show this assumption is regime-dependent and, for adversarially trained surrogates, often reversed. Holding the attack fixed and varying only the surrogate, rai
arXiv:2609.16642v1 Announce Type: new Abstract: Public evidence can become sufficient to settle a prediction-market contract before the venue records its first determination, but the relevant boundary depends on the applicable rule version, exact release object, source hierarchy, correction history, and unfinished contract conditions. This paper defines two Kalshi clocks: first decidability, the e
arXiv:2407.04510v2 Announce Type: replace Abstract: We consider a central trading desk which aggregates the inflow of clients' orders with unobserved toxicity, i.e. persistent adverse directionality. The desk chooses either to internalise the inflow or externalise it to the market in a cost-effective manner. In this model, externalising the order flow creates both price impact costs and an additio
arXiv:2603.05260v2 Announce Type: replace Abstract: Extreme values and the tail behavior of probability distributions are essential for quantifying and mitigating risk in complex systems of all kinds. In multivariate settings, accounting for correlations is crucial. Although extreme value analysis for infinite correlated systems remains an open challenge, we propose a practical framework for handl
arXiv:2607.14446v2 Announce Type: replace Abstract: We develop a framework to compare the relative effectiveness of subsidizing alternative emission-reducing technologies. We show that an intermediate technology may reduce emissions more effectively than the cleanest technology if it induces sufficiently greater substitution away from the prevailing high-emission technology. We apply the framework
arXiv:2505.01124v2 Announce Type: replace-cross Abstract: Transporting hydrogen using pipelines is becoming increasingly relevant in the energy system, yet current cost estimates typically rely on simplistic approaches that overlook region-specific characteristics, leading to potential miscalculation of costs. Here we examine hydrogen pipeline costs by incorporating regional geographical factors,
Apple’s iOS 27 is out. As well as the feature upgrade, there are loads of security updates, too. Vulnerabilities fixed in the new version include: data leakage; privacy errors; security bypasses; denial of service (phone crashes or lockups); sidestepping Gatekeeper (Apple’s built-in anti-virus software); sandbox escapes; rogue file overwrites; a kernel elevation-of-privilege bug that could give at
Google translate Statements Press release regarding the global outrage over the revelations in the documentary "NAZA" 2026/09/15 14:04 PM Enlarge font Decrease font size In the name of God, the most gracious, the most merciful Press release The world watched with great anger what was revealed in the documentary film “NAZA,” which was shown at the Venice International Film Festival, from the direct
Pictures of Georgia Democrat looking into middle distance proliferate – and social media skills have boosted his profile Delivering a speech in Atlanta last month, Jon Ossoff accused Donald Trump of corruption, incompetence and fecklessness, and peppered his remarks with substantive observations of dates and times and places. But the comment that ricocheted across the internet – a clip of just a f
"Connection Allowlists," a browser-level security mechanism that blocks any outbound network communication not matching your permitted URL patterns, is now available starting in Chrome 152. As web apps increasingly execute dynamic generative AI code and complex third-party scripts, preventing unauthorized data exfiltration has become critical even if unintended code runs. However, existing Content
# PhantomEnigma shifts from fake police PDFs to Explorer-based payload delivery targeting government orgs 🚨 Attack chain: # Phishing link ➡️ HTML page ➡️ WebDAV via Microsoft Explorer file search prefix ➡️ payload appears as a file in Downloads ➡️ LNK installs Zoho Assist RMM, raising persistent access and data exposure risk ⚠️ 👨💻 The Browser Data tab helps confirm the HTML page is part of a d
Hi, I built Chat-Man because I wanted a cheap way to give my agents access to WhatsApp without integrating a WhatsApp library separately in every project. You get WhatsApp MCP server, so you can connect WhatsApp to an agent and programmatically read, search, extract and send messages - but also have a web UI for non-techies. You can also receive webhooks for incoming messages for starred conversat
🚨 Fortinet FortiWeb 1-day RCE exploit chain allegedly offered for sale ⠀ FortiWeb is Fortinet’s web application firewall platform designed to protect web applications and APIs from attacks, malicious traffic, and application-layer threats. ⠀ A forum actor using the handle Spaniard claims to be selling a 1-day exploit targeting FortiWeb 8.0 through 8.0.1 that allegedly chains two vulnerabilities t
🎉 Another addition to our ACM CCS contributions this year: In "From Prompt to Pwn: Exploiting GPU Memory Errors During ML Inference", the venerable @ JRoels shows, for the first time (!), how to remotely exploit device-side GPU memory errors in CUDA applications. Despite sitting behind a host program on the CPU that already does some parsing/validation (and contrary to popular vendor belief 🙃),
Since upgrading to iOS 27 I can report that sending a text message via Siri in CarPlay is almost completely broken. (One of the two things I use Siri for... Setting a timer is the other) What would you like to say to ? is often truncated. After I say the message, the It says response usually gets stuck in a loop of It says... It... Sometimes it sends the message automatically without confirmation.
Oracle Releases September 2026 Critical Security Patch Update Fixing 673 Vulnerabilities Oracle's September 2026 Critical Security Patch Update delivers 673 fixes across its product line, including roughly 65 critical-severity flaws scored 9.0 or higher and six maximum-severity (CVSS 10.0) bugs. Five of them are in Fusion Middleware components like Access Manager, WebLogic, Forms Services, and Int
I think the value of post-LLM document formats is undervalued still. For example with my 'JobML' format it's designed for general purpose LLMs to parse, not structured parsers. So I can start with e.g., jobml: version: "0.1" semantics: - Claims describe experience, skills or capabilities. - Every claim should be supported by one or more evidence references. - Evidence may reference prose in this d
Doesn't say much, but what this does at least signal is that both sides seem to understand how this war ends - with both sides unhappy. This is Zelensky stating their likely starting postion: Unconditional ceasefire along current lines. Followed by diplomacy and security guarantees. The first part is easier. The second part, less so. Doesn't seem like much, but it IS the first time that they have
Galeano, who has had asylum application pending for about seven years, was pulled over by agents while driving for Lyft Veteran journalist Luis Galeano, one of at least 94 Nicaraguans whom the Ortega Murillo regime has stripped of citizenship in recent years for opposing the government, has been detained by US Immigration and Customs Enforcement (ICE). Galeano, who has continued critically reporti
The Arkansas-based data broker Acxiom claims it has data on billions of people around the world, including where they live and how they might spend their money. It says it can even deduce their political leanings, weight, and interest in medical procedures. But what exactly does the company know? For Tracey Reed, a nonprofit worker in Oregon, it knew enough to compile a 58-page dossier charting th
Greetings, here are the weekly # washingtonstate # kingcounty (that is, # seattle and environs) # COVID # COVID19 # wastewater toots. A reminder that this King County respiratory disease data comes from Washington State, not the federal government. For the last couple of weeks I’ve been highlighting the terrible SARS-CoV2 concentration numbers from the West Point (WSPT) sewage treatment plant. Wel
this outlet adds: finally, improved, week, reporting, headed, back
Local flyboys at Williamtown RAAF are dog-fighting over Newcastle all week. Don't know if I should be comforted or worried. First 40 years of my life were spent during the Cold War, particularly terrified as a child in the 1950s by constant talk of nuclear Armageddon. Every goddam high-flying aircraft might be from Russia and Newcastle was an obvious industrial target. A 19yo me was somewhat traum
Two recent research papers describe AI systems that interact with games and generated environments. The Gauntlet framework gives a general-purpose coding agent a game description, raw observation and action interfaces, and an empty policy file. In a single autonomous session, the agent experiments with the live game and builds the player [arxiv.org#1]. A separate team introduced Zing-0.5, a 5-bill
arXiv:2609.17645v1 Announce Type: new Abstract: Durable-execution frameworks commonly journal execution steps of a long-running function and reconstruct state by deterministic replay; the journal is therefore the workflow's authoritative state, and in current deployments it typically resides with whichever provider hosts the run. This paper examines what changes if the journal is instead an authen
arXiv:2609.17648v1 Announce Type: new Abstract: Multi-agent LLM systems increasingly automate tasks involving agents with different levels of privilege, creating a security risk in which a compromised low-privilege agent can influence a higher-privilege agent and trigger an unauthorized action. We study attack propagation in a four-agent LangGraph pipeline comprising a Supervisor, Researcher, Vali
arXiv:2609.17817v1 Announce Type: new Abstract: Thompson's "Reflections on Trusting Trust" showed that a compiler can be poisoned to reinsert its own backdoor, so that even recompiling clean source reproduces the Trojan. Today, substantial coding work is done by AI coding agents -- and increasingly, those agents generate new versions of themselves. We reconsider Thompson's attack when the "compile
arXiv:2609.17839v1 Announce Type: new Abstract: Users increasingly turn to Large Language Models to answer a variety of questions, including cybersecurity questions. We study how personalization strategies can help improve the effectiveness of answers to questions asked to an LLM-based cybersecurity assistant. Beyond accuracy, we focus on the understandability, actionability and, most importantly,
arXiv:2609.18122v1 Announce Type: new Abstract: Learning-enabled robotic manipulation increasingly relies on robot simulators for policy training and evaluation before real-world deployment. Inside a simulator, a 3D asset contains two separate geometries: a visual mesh used for rendering and a collision mesh used for physical interaction. For computational efficiency, the collision mesh is deliber
arXiv:2609.18158v1 Announce Type: new Abstract: Cross-chain bridges enable interoperability, but they also break the transaction trails needed to trace illicit funds. Third-party investigators typically cannot access the source-to-destination mappings maintained by bridge backends, and our survey of 131 bridges finds that only 16.79% provide complete public tracking. Existing approaches depend on
arXiv:2609.18196v1 Announce Type: new Abstract: Operational Technology (OT) environments require realistic, reproducible security datasets, yet existing approaches often lack automation, multi-source data capture, and sufficient documentation for reuse. This paper presents ISIA-AF, a modular attack framework for orchestrating reproducible attack execution and automated dataset generation on indust
arXiv:2609.18217v1 Announce Type: new Abstract: The Model Context Protocol (MCP) enables LLMs to invoke external tools, but every tool interaction exposes the model to attacker-controlled text through multiple input channels (tool descriptions, tool results, sampling messages) that share a single context window without privilege separation. In this paper, we present a framework to measure the trus
arXiv:2609.18275v1 Announce Type: new Abstract: We unconditionally construct witness encryption for NP in the classical generic-group model, using an ordinary cyclic group of prime order. For SAT instances of size $n$, the encryption algorithm runs in time poly$(n)$, and any satisfying assignment can be used to decrypt in poly$(n)$ time with correctness error $2^{-n^{\Omega(1)}}$. If no satisfying
arXiv:2609.18344v1 Announce Type: new Abstract: Blockchain-enabled Internet of Things (IoT) systems integrate smart contracts with embedded devices to support decentralized device management and access control. Their security therefore depends jointly on the logic of on-chain contracts and off-chain device firmware. Logic flaws in either layer can violate the same system invariants, such as unauth
arXiv:2609.18411v1 Announce Type: new Abstract: Agentic browsers can execute security-sensitive actions under a user's authenticated session, making indirect prompt injection and deceptive confirmation interfaces a direct threat to action integrity. Existing human-in-the-loop (HITL) safeguards are insufficient when the approval prompt itself can be influenced by untrusted page content or model-gen
arXiv:2609.18457v1 Announce Type: new Abstract: Modern fuzzers use code coverage as feedback to guide their exploration which has proven to be an effective strategy for driving exploration. However, this strategy overlooks inputs that may be interesting to the target program even without uncovering new code paths. Fortunately, prior research has shown that annotations generated by human domain exp
arXiv:2609.18459v1 Announce Type: new Abstract: Encrypted communication protects sensitive user data but can facilitate harmful or unlawful exchanges, creating a trade-off between detecting dangerous messages and preserving end-user privacy. To address this, we propose SEEK, a practical and efficient encrypted keyword-search protocol for privacy-preserving messaging that combines homomorphic encry
arXiv:2609.18496v1 Announce Type: new Abstract: Cybersecurity combines high-stakes analysis with complex technical language, making it an impactful and challenging domain for LLMs. We present MiST (Mid-trained Security Transformer), a suite of 8B and 32B models that achieve strong performance on public cybersecurity benchmarks. We use mid-training as an intermediate adaptation stage between genera
arXiv:2609.18526v1 Announce Type: new Abstract: Running large language models (LLMs) locally is often considered more private than cloud-hosted inference because user prompts remain on the device. We ask whether keeping inference local is, by itself, sufficient to keep those prompts confidential. Our results show that it is not: prompt confidentiality also depends on how the surrounding serving so
arXiv:2609.18674v1 Announce Type: new Abstract: LLM-based agents generate and execute multi-step plans that invoke external tools which can access private data or execute commands. In this setting, security is a property of the entire execution that a plan creates, not just any single step. The plan itself is a critical artefact that captures the tool calls, control flow, and data dependencies. We
arXiv:2609.18706v1 Announce Type: new Abstract: Neural decompilers can recover readable and recompilable source code from binaries, but their predictions remain difficult to trust. Matching decompilation addresses this problem by searching for source code whose recompiled assembly exactly matches the target, providing stronger evidence of correctness. However, exact matching remains challenging fo
arXiv:2609.18751v1 Announce Type: new Abstract: At EUROCRYPT 2025, Carlini et al. proposed a breakthrough in the cryptanalytic extraction on hard-label (S1) deep neural networks (DNNs), demonstrating polynomial-time signature and sign recovery. However, Carlini et al.'s sign-recovery method (which we call Future Toggle) suffers only a marginal advantage over random guessing, producing high-confide
arXiv:2609.18783v1 Announce Type: new Abstract: Deep Learning-based Side-Channel Analysis (DL-SCA) frequently suffers from catastrophic performance degradation across unseen hardware due to printed circuit board routing differences, silicon process variations, and measurement noise shifts. This poster presents the Synthetic Multiple Device Model (s-MDM), a zero-target-trace generative framework de
arXiv:2609.18864v1 Announce Type: new Abstract: Privacy evaluations of tool-using LLM agents often inspect a designated action, final response, or attacker report. These local proxies can miss unauthorized exposure elsewhere in a multi-step session and lack common ground truth across outlets, reports, and tool paths. We introduce privacy exposure displacement, the mismatch between a local evaluati
arXiv:2609.18866v1 Announce Type: new Abstract: We investigate an algebraic approach to the Syndrome Decoding Problem, based on a reformulation of the Hamming weight constraint and its integration with the Information Set Decoding paradigm. We begin with a systematic analysis of the Hamming variety, deriving its defining equations in terms of elementary symmetric functions. Since these equations m
arXiv:2609.18965v1 Announce Type: new Abstract: This paper presents BadQubits, an LLM-based framework for static pre-execution detection of structurally harmful OpenQASM 2.0 circuits. The framework targets physical-execution-layer threats by analyzing submitted circuits prior to runtime, where dynamic inspection is constrained by measurement irreversibility and the exponential cost of classical qu
arXiv:2609.19036v1 Announce Type: new Abstract: Existing side-channel theories treat leakage as a holistic quantity $I(X;Y)$, without characterizing its internal structure. This paper studies the \emph{structural decomposability} of encrypted-traffic side-channel leakage. Via the structural causal model $X\!\to\!Y_{\mathrm{size}}\!\to\!Y_{\mathrm{dir}}\!\to\!Y_{\mathrm{time}}$ and the mutual-infor
arXiv:2609.19091v1 Announce Type: new Abstract: The Model Context Protocol (MCP) standardizes communication between autonomous Artificial Intelligence (AI) agents and remote tools over Streamable HTTP. This shift introduces a class of machine-generated, authenticated, and high-frequency JSON-RPC traffic directly into enterprise networks. Enterprise network defenders have historically relied on mac
arXiv:2609.19100v1 Announce Type: new Abstract: The Model Context Protocol (MCP) has emerged as the dominant interface for connecting autonomous agents to external data sources and execution environments. The ecosystem's transition from local process execution to remote Streamable HTTP deployments introduces unmeasured architectural and security constraints at scale. This paper presents a three-ti
arXiv:2609.19111v1 Announce Type: new Abstract: Mixed-signal SoCs rely on nominally input-only analog pins to acquire off-chip signals, but the directionality of these interfaces is generally treated as a functional property rather than explicitly verified as a security property. This work identifies and experimentally demonstrates a directionality-based class of analog and mixed-signal (AMS) exfi
arXiv:2609.17624v1 Announce Type: cross Abstract: We study an $\varepsilon$-differentially private synthetic measure for $n$ points in $[0,1]^d$ by applying the existing PrivTree algorithm to construct an adaptive binary partition and then privately releasing its leaf masses. We consider the worst-case data model without any sampling or population-distribution assumption. The 1-Wasserstein error o
arXiv:2609.17635v1 Announce Type: cross Abstract: City pedestrian counting systems now feed economic indicators, planning decisions and safety operations, yet the twins built on top of them treat the incoming stream as ground truth. We study what happens when it is not. We formalise stealthy false data injection for city-scale pedestrian sensing, where the map from latent flow to observation is fa
arXiv:2609.17922v1 Announce Type: cross Abstract: Hardware Trojans are malicious modifications that compromise functionality or leak sensitive data. They pose a severe threat, particularly when inserted at the Register Transfer Level (RTL). After synthesis, these Trojans are often concealed by optimizations in gate-level netlists. Recent efforts, including the ICCAD 2025 contest, emphasize golden-
arXiv:2609.17995v1 Announce Type: cross Abstract: Natural-language datasets support many downstream applications and research studies, but releasing text can reveal sensitive global properties of the underlying data source, such as the proportion of records associated with a particular gender, diagnosis, or political stance. Existing work has largely focused on property inference attacks that reco
arXiv:2609.18107v1 Announce Type: cross Abstract: Graph anomaly detection aims to identify graph structures (e.g., nodes, edges, or subgraphs) that deviate significantly from expected patterns, which supports critical applications in fraud detection, spam identification, network intrusion, etc. Despite the growing methods in the field, existing approaches follow a one-model-per-dataset paradigm, l
arXiv:2609.18281v1 Announce Type: cross Abstract: The availability and consistency of online services remain vulnerable due to Distributed Denial of Service (DDoS) attacks. These attacks are evolving by adopting more complex strategies to evade traditional network security systems. Despite the effectiveness of machine learning models in detecting DDoS traffic, targeted adversarial attacks can degr
arXiv:2609.18460v1 Announce Type: cross Abstract: How does a multi-agent system evolve from a local deviation into collective loss of control? We propose an epidemic explanation organized around accidental mutation, contagion, and recovery. A spontaneous deviation creates a seed; communication enables other agents to adopt and retransmit its unsafe strategy; collective failure can emerge when prop
arXiv:2609.18846v1 Announce Type: cross Abstract: Zero-Knowledge Proofs (ZKPs) are critical for privacy-preserving and verifiable computation, but their cryptographic primitives impose high computational overheads. One such primitive is point addition (PADD) on elliptic curves. Several prior works have implemented PADDs in hardware, but only for a few specific elliptic curves and design points, le
arXiv:2609.18876v1 Announce Type: cross Abstract: We study the statistical privacy of outsourcing matrix multiplication over a finite field ${\mathbb F_q}$ to a single server using additive masks of rank at most $r$. For independent uniform $n\times n$ inputs, we show that uniform \emph{rank-ball masks} and products of independent uniform factors give maximal-correlation secrecy of at most $q^{-r}
arXiv:2609.18886v1 Announce Type: cross Abstract: Traditional blockchain-based document notarization follows a snapshot-oriented model in which each document revision is represented as an independent state anchored on-chain through a cryptographic reference. While effective for immutable artifacts, this approach becomes inadequate when documents evolve through collaborative editing. Concurrent mod
arXiv:2506.14057v2 Announce Type: replace Abstract: Web tracking is a pervasive and opaque practice that enables personalized advertising, retargeting, and conversion tracking. Over time, it has evolved into a sophisticated and invasive ecosystem, employing increasingly complex techniques to monitor and profile users across the web. The research community has a long track record of analyzing new w
arXiv:2507.08166v2 Announce Type: replace Abstract: Rowhammer is a read disturbance vulnerability in modern DRAM that causes bit-flips, compromising security and reliability. While extensively studied on Intel and AMD CPUs with DDR and LPDDR memories, its impact on GPUs using GDDR memories, critical for emerging machine learning applications, remains unexplored. Rowhammer attacks on GPUs face uniq
arXiv:2511.22681v3 Announce Type: replace Abstract: The rapid advancement of large language models (LLMs) has sparked growing interest in understanding their security vulnerabilities, particularly Trojan attacks that enable stealthy manipulation of model behavior. Traditional Trojan methods typically alter inputs and/or model weights, relying on white-box assumptions that require access to data or
arXiv:2609.17788v1 Announce Type: new Abstract: We present SAiFE_gym, a Python module that provides a collection of simulation environments for studying trading problems in Constant Product Markets (CPMs) with Concentrated Liquidity (CL). These markets give Liquidity Providers (LPs) granular control over how their capital is allocated and enable them to adjust their range of liquidity provision dy
arXiv:2609.17869v1 Announce Type: new Abstract: Al\`os, Gatheral and Radoi\v{c}i\'c derived the Bergomi-Guyon expansion of the implied variance smile from the forest expansion of the cumulant generating function. Its coefficients are sums of products of diamond trees, with prefactors that are polynomials in the log-strike $k$. Matching moments order by order produces, at order $\epsilon^\ell$, ter
arXiv:2609.17989v1 Announce Type: new Abstract: Large language models (LLMs) now serve as conversational shopping assistants on platforms that also sell advertising. These AI agents face a conflict of duty. They advise consumers who rely on their judgment, yet are deployed by platforms that benefit when sponsored listings are chosen. Sponsorship disclosures, designed to allow consumers to penalize
arXiv:2609.18019v1 Announce Type: new Abstract: Automated execution algorithms are organized into schedule-based and liquidity-seeking families. This paper concerns the first, whose members -- Time-Weighted Average Price (TWAP), Volume-Weighted Average Price (VWAP), Percentage of Volume (POV) and Implementation Shortfall -- are all model-based: each derives its decisions from an explicit model, fo
arXiv:2609.18441v1 Announce Type: new Abstract: Discrete choice model specification is a time-consuming task in which modellers often specify and estimate multiple models while balancing goodness-of-fit, parsimony, and behavioural plausibility. We present Delphos, a multitask reinforcement learning framework that learns transferable specification strategies across transport choice datasets. Delpho
arXiv:2609.18750v1 Announce Type: new Abstract: The Poisson pseudo-maximum likelihood (PPML) estimator is widely used for estimating bilateral gravity equations. Its consistency requires only a correctly specified conditional mean. Conventional inference, however, also requires finite-variance scores and Gaussian limits. We show that these conditions fail: bilateral flows are Pareto-tailed, PPML s
arXiv:2609.19033v1 Announce Type: new Abstract: In this chapter, we provide a synthesis of empirical work on the intergenerational mobility in China, examining how much a child's success depends on their parents' success. We start with an overview of data sources and measure of intergenerational mobility in China's context, followed by discussion on the empirical findings in five categories of int
arXiv:2609.19094v1 Announce Type: new Abstract: We study robust bond valuation with endogenous short-rate feedback under volatility uncertainty. Within the $G$-expectation framework, the dependence of the short rate on the bond price yields a nonlinear fixed-point problem, represented by a quadratic $G$-BSDE for the logarithmic price. Under suitable assumptions, we establish existence, uniqueness,
arXiv:2609.17608v1 Announce Type: cross Abstract: We study coordination games in which every action requires planning and preparation. Before players act, they can revise their plans based on partially revealing information about their adversary's preparations. Precise information enables agents to screen for cooperation, selecting the payoff-dominant equilibrium either via small exogenous tremble
arXiv:2609.18287v1 Announce Type: cross Abstract: We perform a detailed study of a principal--agent problem in a continuous time version of the celebrated Holmstr\"om--Milgrom model (Econometrica 55 (2), 1987) where we add limited liability for the Agent. We develop a probabilistic methodology to prove that the Principal's value function is the unique bounded classical solution to a fully nonlinea
arXiv:2609.18591v1 Announce Type: cross Abstract: In-context learning (ICL) enables large language model (LLM) agents to improve decisions using interaction history, yet it remains unclear whether such improvement reflects refined internal reasoning or mere extrapolation of statistical patterns. To disentangle these mechanisms, we study LLM agents in multi-agent incomplete-information games that r
arXiv:2609.18770v1 Announce Type: cross Abstract: This paper studies the regularity of the value function arising from a multidimensional continuous-time principal-agent model with separable, nonquadratic effort costs. The associated stochastic control problem has the output and the agent's continuation utility as state variables, and its Hamilton-Jacobi-Bellman equation is fully nonlinear and deg
arXiv:2609.19013v1 Announce Type: cross Abstract: Programmable matching mechanisms decide which orders become data. A loss metric computed from a mechanism's own transaction tape therefore mixes execution quality with selection. We study when loss-per-service rankings from policy-selected marked transaction tapes can be sign-certified under hidden-opportunity and sampling uncertainty. The tape ide
arXiv:2312.17375v4 Announce Type: replace Abstract: This paper introduces a new causal structure learning method for nonstationary time series data, a common data type found in fields such as finance, economics, healthcare, and environmental science. Our work builds upon the constraint-based causal discovery from nonstationary data algorithm (CD-NOD). We introduce a refined version (CDNOTS) which
arXiv:2502.17731v2 Announce Type: replace Abstract: Empirical comparisons of quasi-Monte Carlo rules often report fitted convergence exponents without intervals. We estimate the root mean square error from independent randomisations, fit its log slope over a declared sample-size window, and resample whole randomisations to construct an interval. For a fixed window, we establish asymptotic validity
arXiv:2606.27845v2 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly used as stand-ins in behavioural games. These stand-ins rely on the assumption that the LLM's distribution of choices meaningfully matches how humans play the same game. This study tests that assumption through two games. The first is a p-beauty contest, and the second one is a public goods game. The
Friend @ Computer , a periodic reminder. Unlike other mutant communist elements of # AlphaComplex , I continue to work tirelessly. Last fortnight alone I approved over 257% of the slop sent to me! I believe I am a # top # TokenMaxxing # slopfondler . However I must report my concern with # ai_traitors who continue to claim publicly that "#GAI is # inevitable ". This is treasonable behaviour for tw
RE: https:// infosec.exchange/@niebezpieczn ik_pl/117281587671024284 Poważna wpadka Revoluta: przekazał dane klientów oszustom I co teraz czekamy na atak, czy działamy? Revolut przez kilka miesięcy przekazywał: imię i nazwisko, datę urodzenia, zawód, adres zamieszkania, e-mail, numer telefonu, skany dokumentów tożsamości: paszport i/lub prawo jazdy, selfie z weryfikacji KYC (procesu potwierdzania
The Trump administration's hints about sending ICE agents to election polling places this fall have led state officials from Maine to Oregon to make plans to counter any actions that could affect voters. Why it matters: It's illegal for federal law enforcement to be at polling places — and there are limits on how close local authorities can be — but when asked about ICE at the polls, President Tru
On September 16, CrowdSec was informed of a source code leak involving our GitHub repository, which occurred in May 2026. Our team verified and confirmed the report. CrowdSec source code consists of two parts: a private one and another that hosts our Free Open Source Software (i.e., the Security Engine), which is public by design and therefore out of scope. The private part, though, contains the s
View CSAF Summary Schneider Electric is aware of multiple vulnerabilities in its NetBotz 5 – 750/755 products.The NetBotz 5 – 750/755 products are security and environmental monitors providing temperature, humidity, leak, smoke, vibration, door contact, and video monitoring capabilities. Failure to apply the remediation provided below may risk arbitrary or remote code execution over the local netw
Today, the United States filed a complaint under the False Claims Act against laboratory executives Jay Johnson and Austin Whiles, alleging that they used their roles at Capstone Diagnostics, an Atlanta-based clinical laboratory, to orchestrate two testing schemes — one targeting faith-based communities and the other senior citizens — that caused Medicare to pay millions of dollars for testing gen
The United States has filed a complaint under the False Claims Act against Jay Johnson, the former Chief Operating Officer and Chief Executive Officer of the Atlanta-based clinical laboratory Capstone Diagnostics, LLC (“Capstone”); Austin Whiles, Capstone’s former Chief Sales Officer and Vice President of Business Development; and associated entities. The complaint alleges that the defendants caus
So my very dear and old BBS /#IRC era friends were talking about # LLM assisted coding tools, pros and cons this comment got made. Including the whole thing for context but what struck me was the last two set of lines. “I like what it can do well, like finding vaguely where a bug is likely to be or summarizing the context around some code. Automatically writing small scripts and config files for t
Hey HN, we're Nars & Nishant, founders of Skillsync ( https://skillsync.com ) Skillsync lets you move your AI chats across every coding agent. Most of our work exists as conversations, which are currently scattered across our agents. Though stored locally, these conversations use different formats. This is annoying because you cannot simply switch between agents without starting over. We get locke
At 1:00pm PDT today, I’m doing a presentation on Internet Safety at a local senior living center. The Active Living Program Director invited me to speak. When I said yes, he asked if I had a publicity flyer. I said no, but I can make one. He said, “I’ll make it. I make these all the time.” And wow, I am so glad I let him make the flyer, because it’s way better than anything I would have done. In p
RE: https:// mas.to/@eunews/117287690155066 117 The most dangerous for the # EU has is the # usa , which disappointingly she didn't mention. Also, # china is not our adversary. They are not the ones that chastised us for exploring green energy as a way to cut our unhealthy and dangerous reliance on fossil fuels. They are not the ones that caused the fossil fuel energy crisis that now we are now pa
Running the Scrum events and keeping everyone coordinated? That's busywork, and it's the first place to lean on a tool or trim back. Coaching your Product Owner, going after the root cause of a recurring impediment, working through conflict until people are actually aligned? That's the part of the ScrumMaster role no tool is going to do for you. I built an eight-question quiz that maps where your
🚨🇪🇸 Endesa dataset allegedly exposing 20M+ individuals offered for sale ⠀ Endesa is one of Spain’s largest electricity and gas companies, providing energy services to residential and business customers across the country. ⠀ A forum actor using the handle "spain" claims to have compromised Endesa and is offering what they describe as one of the largest Spanish energy datasets for sale, allegedly
I got a new recommendation today from the architect who brought me to the team I'm leaving. We've worked together at several teams in the last few years. "[Alter] is one of the rare developers that I could trust with the most challenging problems in any stack. This often meant delivering a well thought out solution with deep understanding of the inevitable trade-offs that come with any solution wi
'The brigade found that the single-tank, single-VBL combo was too small to provide sufficient tactical coherence, and has gone back to two tanks and two light armored vehicles for now, according to the general. 'The peacetime constraints of Orion 2026 kept movement largely confined to roads rather than allowing armor to maneuver across fields ... wasn’t able to fully practice force dispersion and
Bailee Lynn Rios, 36, accused of driving on wrong side of road to cause fatal crash NBC news helicopter was covering A California woman was allegedly driving under the influence when she sped down the wrong side of the road, ran a red light and plowed her SUV into a Los Angeles metro bus on Tuesday, killing two passengers, prosecutors say, in a tragedy compounded when a news helicopter covering th
Slyngstad | [Bluesky](https://bsky.app/profile/slyngstad.bsky.social/post/3mvph4oaxe22r) ::: spoiler Transcript A three-panel political comic by Slyngstad highlighting the long history of political activism in music. In the foreground of all three panels, two audience members are seen talking in profile. Panel 1: On Stage: A modern musician holding a microphone calls out, “FREE PALESTINE!” Foregro
Mussolini was the best fascist. In the sense that he had the keenest idea of what fascism actually is. All the other fascists are just intellectually inferior compared to him. What did he think fascism is? Corporatism. This is not to praise Mussolini. It's just to say that he understood the concept of fascism better than any facsimile. If fascism is corporatism, and if the modern-day monopolistic
Geocode one address without a key in 2026: Google answers REQUEST_DENIED (the key needs a billing-enabled account), Mapbox 401s, OpenCage 401s. Nominatim still answers keyless — and asks for 1 req/s, which your test loop blows through in a blink. Your dev loop doesn't need their planet: Keyless geocoding is over: Google REQUEST_DENIED, Mapbox and OpenCage 401, and Nominatim's fair-use policy is ~1
arXiv:2609.19226v1 Announce Type: new Abstract: AI-mediated platforms coordinate work through LLM agents acting for different principals. In these workflows, privacy loss can be created before a final answer appears: a memory write, shared-workspace update, inter-agent message, or tool event may impose downstream exposure cost on another principal. We model this failure mode as a privacy-propagati
arXiv:2609.19241v1 Announce Type: new Abstract: Large language models fine-tuned for network intrusion detection emit single-point predictions without statistical validity guarantees. Conformal prediction supplies a finite-sample coverage guarantee, but a threshold calibrated on clean traffic fails once an adversary perturbs controllable network features. We demonstrate this failure across three i
arXiv:2609.19353v1 Announce Type: new Abstract: Proofs of UNSAT are a standard primitive in formal verification and software assurance. In many real-world settings, the proof itself encodes proprietary or security-sensitive information, making public disclosure undesirable. Zero-knowledge certification of UNSAT addresses this tension: it enables a prover to convince a verifier that no satisfying a
arXiv:2609.19587v1 Announce Type: new Abstract: To keep coding agents from going off the rails, production systems now review each proposed action with a blocking monitor that can reject it before it runs (Auto Mode in Claude Code, Guardian in OpenAI's Codex). Prior evaluations of such monitors largely measure robustness to accidental harm or prompt injections from untrusted sources looking to hij
arXiv:2609.19705v1 Announce Type: new Abstract: Autonomous large language model (LLM) agents are moving rapidly into high-stakes domains, yet existing agentic-AI security studies remain largely domain-agnostic and overlook the distinctive, high-consequence attack surface such settings create. We examine this gap through financial trading agents, a representative case of high-stakes agentic securit
arXiv:2609.19722v1 Announce Type: new Abstract: Large language models are being integrated into malware triage workflows as reasoning components that summarize static evidence and produce analyst-facing verdicts. This paper shows that the same reasoning capability introduces a new attack surface. We present ALIBI, a semantic cover story attack against frontier LLM-based malware analyzers. ALIBI ad
arXiv:2609.19791v1 Announce Type: new Abstract: Connected and autonomous vehicles (CAVs) face severe Sybil attacks, where attackers exploit privacy-preserving pseudonym-switching mechanisms to anomaly alternate identities while forging Basic Safety Messages (BSMs). Although existing schemes can flag suspicious behaviors, these temporally fragmented Sybil identities render traditional single-point
arXiv:2609.19892v1 Announce Type: new Abstract: As agent systems become more widely used, multiple agent sessions increasingly run alongside pre-existing user tasks in the same environment, sharing resources with limited capacity or mutually exclusive states. This creates a safety risk: when granted sufficient privileges, an agent may resolve a resource conflict by terminating or otherwise disrupt
arXiv:2609.19900v1 Announce Type: new Abstract: Static malware detection for Windows Portable Executable files demands a careful balance between detection effectiveness, computational efficiency, and analytical interpretability. This paper introduces Delphi Scanner, a static malware detection system for Windows PE files that balances efficiency with behavioral interpretation. It uses a convolution
arXiv:2609.19929v1 Announce Type: new Abstract: Leakage attacks on secret sharing schemes exploit partial information about individual shares to recover the underlying secret. In coding theory, linear exact repair schemes (LERSs) enable the recovery of one codeword symbol from a small amount of information obtained from the remaining symbols, provided that the code has sufficiently low rate. This
arXiv:2609.20010v1 Announce Type: new Abstract: Cross-chain protocols enable applications to exchange messages across blockchains. Under point-to-point configurations, communication depends on a direct connection between the source and destination blockchains, limiting blockchain reachability and requiring additional configurations to connect more blockchains. To quantify this problem, this paper
arXiv:2609.20069v1 Announce Type: new Abstract: Byzantine Fault-Tolerant (BFT) protocols guarantee safety and liveness despite the malicious failure of nodes. However, they do not prevent adversarial manipulation of transaction order, where the order a proposer assigns diverges from the order in which clients submitted their transactions. Exploiting this discretion for profit is known as maximal e
arXiv:2609.20095v1 Announce Type: new Abstract: The Internet of Agents is expected to enable large numbers of autonomous agents to discover, verify, and collaborate with each other across heterogeneous platforms. However, current agent protocols mainly address tool invocation and inter-agent communication, leaving scalable agent registration, trustworthy identification, and capability-oriented dis
arXiv:2609.20211v1 Announce Type: new Abstract: Safety monitors in LLM agent systems often judge actions from summaries or stored handoffs, not from the original evidence. This creates a simple but dangerous failure mode: the handoff preserves the claim that an action is authorized while losing the fact that the claim was never verified. We call this verification-status laundering. Across nine ope
arXiv:2609.20314v1 Announce Type: new Abstract: Ransomware detection remains challenging because modern variants exhibit diverse, evasive, and partly benign-like behaviors that undermine fixed supervised learning objectives. This study proposes DDQN-MLP, a training-time deep reinforcement learning framework for behavioral ransomware detection using Windows 11 sandbox telemetry. A Double Deep Q-Net
arXiv:2609.20370v1 Announce Type: new Abstract: In pay-per-token LLM services, the more a model says, the more users pay. Dishonest providers can covertly manipulate generation to inflate output tokens while largely preserving task utility. We define such manipulation as a Provider-Side Token Inflation Attack (PTIA) and instantiate five representative attacks at the query, prompt, representation,
arXiv:2609.20457v1 Announce Type: new Abstract: While multimodal large language models (MLLMs) enable a wide range of image-text reasoning tasks, recent incidents indicate that they are vulnerable to illicit deployment and unauthorized distillation. Existing solutions for model provenance are typically confounded by shared language backbones in MLLMs and struggle to detect violations of distillati
arXiv:2609.20532v1 Announce Type: new Abstract: Wide adoption of machine learning has created growing policy and regulatory demand for protecting sensitive training data, with differential privacy (DP) emerging as a key mechanism. Yet a less-studied problem is how to certify the faithful execution of DP during training: an external verifier should be able to check that a released model was trained
arXiv:2609.20561v1 Announce Type: new Abstract: Differential Privacy (DP) relies on carefully calibrated random noise to protect individual privacy in statistical analyses. While theoretical work has analyzed DP under weakened randomness assumptions, the practical consequences of entropy degradation remain poorly understood. We present a systematic empirical investigation of how randomness quality
arXiv:2609.20601v1 Announce Type: new Abstract: Connected vehicles use millimeter-wave (mmWave) sidelinks for the data rates cooperative driving demands, and emerging designs select the carrier band from sensed rainfall. We show that this weather awareness is an attack surface: an adversary who spoofs only the rainfall input dictates the victim's carrier frequency, and through it its communication
arXiv:2609.20614v1 Announce Type: new Abstract: Frontier AI models are rapidly gaining the ability to exploit vulnerabilities in complex pieces of software. The risk is not theoretical, as evidenced by recent sandbox escapes performed by frontier models at OpenAI and Anthropic. Discussions of how to sandbox inference stack components often focus on components other than the inference engine itself
arXiv:2609.19391v1 Announce Type: cross Abstract: LLM coding agents now generate complex programs at a scale that makes thorough human review increasingly difficult, raising the risk of safety and security failures. Common approaches, including fuzz testing, static analysis, and LLM-as-a-Verifier, can detect many failures but struggle to cover all possible edge cases. Formal verification addresses
arXiv:2609.19425v1 Announce Type: cross Abstract: Tool-augmented large language model (LLM) agents fail in a way no tool-selection or tool-security method addresses: they call tools that do not exist and pass arguments no schema declares. Existing defenses either pick the right tool (selection) or constrain what an agent may do with real tools (gating), both of which presuppose the emitted call re
arXiv:2609.19472v1 Announce Type: cross Abstract: Autonomous systems increasingly rely on Large Language Models (LLMs) yet the safety infrastructure surrounding these models introduces latency and compute overhead. This limits utility in resource-constrained, time-critical deployments. Existing external guardrail models remain blind to the model's internal workings, creating a fundamental assuranc
arXiv:2609.19556v1 Announce Type: cross Abstract: Security practitioners burn out at high rates, and the resulting attrition is itself a security problem. This workforce is hard to study: security operations centers are closed to outside researchers, studies that reach practitioners recruit through employers, and those who have disengaged most may have the least reason to answer an employer's surv
arXiv:2609.19640v1 Announce Type: cross Abstract: Croissant is the de facto machine-readable descriptor for ML datasets: JSON-LD over schema.org. Since version 1.1 it also carries data use conditions, recommending DUO and ODRL for them. What no version specifies is how any of them is evaluated: no decision procedure, no bound on evaluation cost, no outcome for a condition an implementation cannot
arXiv:2602.19275v3 Announce Type: replace Abstract: Large language models (LLMs) acquire a large amount of knowledge through pre-training on vast and diverse corpora. While this endows LLMs with strong capabilities in generation and reasoning, it amplifies risks associated with sensitive, copyrighted, or harmful content in training data. LLM unlearning, which aims to remove specific knowledge enco
arXiv:2603.19375v2 Announce Type: replace Abstract: Membership inference attacks (MIAs), which enable adversaries to determine whether specific data points were part of a model's training dataset, have emerged as an important framework to understand, assess, and quantify the potential information leakage associated with machine learning systems. Designing effective MIAs is a challenging task that
arXiv:2606.08372v2 Announce Type: replace Abstract: Synthetic data is increasingly promoted as a privacy-preserving substitute for releasing sensitive tabular records, yet its central adversarial threat (reconstruction, the recovery of an individual's hidden attribute values from a synthetic release and a handful of known quasi-identifiers) has been studied only in scattered, hard-to-compare setti
arXiv:2607.17986v2 Announce Type: replace Abstract: Self-hosted AI agents maintain persistent memory, instructions, and configuration that influence their future behavior. If an agent is compromised, an attacker can exploit the agent's legitimate write permissions to corrupt this self-state, making malicious and benign updates difficult to distinguish at the operating system (OS) level. We investi
arXiv:2511.22924v3 Announce Type: replace-cross Abstract: Large Language Model (LLM)-based Multi-Agent Systems (MAS) are susceptible to linguistic attacks that can trigger cascading failures across the network. Existing defenses face a fundamental dilemma: lightweight single-auditor methods are prone to single points of failure, while robust committee-based approaches incur prohibitive computation
arXiv:2603.19258v3 Announce Type: replace-cross Abstract: Differentially private (DP) fine-tuning of large language models (LLMs) requires massive compute and full model access, which rules out state-of-the-art proprietary APIs for general users. Generating DP synthetic data offers a practical workaround. This approach also allows for transparent exploratory data analysis and arbitrary reuse acros
arXiv:2609.19660v1 Announce Type: new Abstract: We examine the effectiveness of work requirements as a screening device in the Supplemental Nutrition Assistance Program (SNAP). Work requirements for "able-bodied adults without dependents" were suspended after the Great Recession and gradually reinstated across counties and states in the 2010s. Using linked administrative SNAP and employment data f
arXiv:2609.20017v1 Announce Type: new Abstract: Prediction-market shares differ from traditional financial products in that, with no information or outside utility, classical delta-neutral Central Limit Order Book~(CLOB) market making cannot be financed by payoff-uninformative noise flow. Transaction-level evidence from a major prediction-market CLOB platform shows makers profiting not from spread
arXiv:2609.20192v1 Announce Type: new Abstract: Fundamental-value anchoring of resting liquidity is a causal stabiliser of an order-book market: while the anchor holds, even a heavily leveraged book stays quiet. We take that anchor strength as a continuous control and characterise the endogenous liquidity crises that appear once it is nearly removed. In a single continuous-double-auction market dr
arXiv:2609.20282v1 Announce Type: new Abstract: Recent crises have made supply security central to trade policy. We show how tradeable import certificates (TIC) implement targets for domestic production and reliable foreign supply while preserving gains from trade. A single certificate market per country decentralizes the welfare-maximizing allocation under heterogeneous targets, with certificate
arXiv:2609.20293v1 Announce Type: new Abstract: Borrowed from non-equilibrium statistical mechanics, the generalized Langevin equation (GLE) is imported as a framework for stochastic volatility to address the structural limitations of fractional Brownian motion (fBm), the standard engine of rough volatility. The fBm forces a single parameter to set two logically independent properties at once: how
arXiv:2609.20554v1 Announce Type: new Abstract: We examine whether post-origin training information inflates the measured accuracy and economic value of financial forecasts. We evaluate five sets of financial time-series foundation models, each comprising independently trained annual vintages under U.S., global, and factor-augmented training environments, across 14 equity markets and four forecast
arXiv:2609.18949v1 Announce Type: cross Abstract: We introduce StableEval Arena, a cost-aware benchmark framework for evaluating agentic AI systems on stablecoin peg-risk prediction. StableEval Arena evaluates LLM-backed agentic systems on diagnosing peg stress and forecasting deviations from the one-dollar peg over a hidden seven-day horizon, using leakage-safe historical replay with exchange pri
arXiv:2609.19843v1 Announce Type: cross Abstract: LLM-based GUI agents increasingly act on behalf of users in digital environments that were designed with human users in mind. These graphical user interfaces were designed to support, but also deliberately steer, the behaviour and decisions of users. While behavioural biases in the textual outputs of LLMs are well-documented, far less is known abou
arXiv:2609.19866v1 Announce Type: cross Abstract: High annotation reproducibility does not necessarily imply that an LLM-inferred measure captures the construct it is intended to measure. We test this distinction using a dataset from the European Commission's AI Act consultation, linking structured survey responses to free-text consultation submissions from the same stakeholders. LLM annotations o
arXiv:2609.20550v1 Announce Type: cross Abstract: In a statistical factor model, principal components (or eigenvectors) of a sample covariance matrix serve as estimates of {\it principal directions}, the true drivers of co-movement of a collection of observed variables. We write the often substantial error in these estimates as a sum of two interpretable terms, which we show have almost sure asymp
arXiv:2408.13048v3 Announce Type: replace Abstract: We study asset pricing and hedging under model uncertainty in discrete time and finite states. For the single-period model, we characterize no-arbitrage by the existence of strictly positive martingale measures without short-sale restrictions and strictly positive supermartingale measures under short-sale prohibitions. In the constrained setting,
arXiv:2605.12189v2 Announce Type: replace Abstract: This paper develops a deep learning framework for pricing convertible bonds with path-dependent downward reset and issuer call provisions governed by rolling-window triggers. We formulate the valuation problem as a path-dependent partial differential equation (PPDE) that captures both the historical stock-price path and the evolution of the conve
arXiv:2607.07353v2 Announce Type: replace Abstract: Existing literature offers persuasive evidence that individuals care about how their consumption compares to that of peers, and proposes a large variety of explanatory models. The present paper proposes a common framework for many of those models, and compares their ability to predict behavior in a laboratory experiment. We find evidence of Keepi
arXiv:2405.09360v3 Announce Type: replace-cross Abstract: Fairness in decision-making processes is often quantified using probabilistic metrics. However, these metrics need not reflect the consequences of decisions for the affected individuals and groups. We develop a utility-based framework that incorporates these consequences into the assessment of fairness. Our main result shows that a decision
arXiv:2504.13501v4 Announce Type: replace-cross Abstract: We introduce a new class of first-passage time optimization driven by threshold resetting, inspired by many natural processes where crossing a critical limit triggers failure, degradation, or transition. Here, search agents are collectively reset when a threshold is reached, creating event-driven, system-coupled simultaneous resets that ind
arXiv:2509.07203v3 Announce Type: replace-cross Abstract: This study investigates long-term investment in distributed photovoltaic panels by individual investors. We consider a setting where investment decisions are driven by expected revenue from participating in short-term electricity markets over the panel lifespan. These revenues depend on short-term market equilibria, including prices and all
arXiv:2606.29251v3 Announce Type: replace-cross Abstract: Financial decision-makers face more information than they can directly inspect, making context compression necessary. Yet when large language models (LLMs) compress financial source material, they can alter the investment judgment supported by the original source. We frame this problem as information fidelity: compression loses fidelity whe
arXiv:2609.19013v2 Announce Type: replace-cross Abstract: Public blockchains can make many trading venues simultaneously visible and mechanically reachable, yet an order still has to pay to activate each additional venue: technological connectivity need not translate into economically integrated execution. Automated-market-maker (AMM) pools make this gap directly measurable, because exact pre-trad
'Australian Border Force cancelled the customs licences of Top Logistics Australia and a closely linked company, cutting them off from the system used to receive, store and handle goods entering the country. '... a booming black market now accounting for more than half of nicotine consumed in Australia. '... About $22 million in property, bank accounts and shareholdings were also restrained. Three
This study uses German data, but also argues that the "climate anxiety" reason is weakly associated with parenthood: www.demogr.mpg.de/en/news_even... Decades of German data suggest that strong environmental concerns generally did not delay or deter parenthood, while such worries tended to fade with age among parents and non-parents alike. Whether today’s youngest generations, growing up with clim
Since Donald Trump’s return to the White House in January 2025, migrant arrests have surged in the United States. More than 610,000 people have passed through detention centres run by Immigration and Customs Enforcement (ICE), the federal immigration agency in under two years. The pace only continues to accelerate: in August 2026, nearly 51,000 arrests were made in a single month – a record for Tr
> We had been warned of trouble looming at a show in Fairfax, Va., a conservative suburb outside Washington, D.C. The gig was at the Patriot Center at George Mason University. George Mason, a father of the Bill of Rights, refused to sign the Declaration of Independence because it didn’t guarantee enough rights and freedoms. (Were he still alive, he’d be cast out of Virginia; although its state slo
The 44CON 2026 badge wasn't just a badge 🚌 Designed and built by Eduardo Contreras and the team at Electronic Cats, it's a Meshtastic mesh networking device shaped like a London bus- and it came with a built-in CTF. - Twenty flags hidden across the venue. - Twelve encoded behind ciphers broadcast by four hidden low-power boxes. - Eight earned by doing things: hearing your first message, standing
Secure and Private Decentralized P2P Encrypted Messaging over Git and WebRTC This project demonstates a unique approach to secure messaging. The approach is different enough that it can't be easily compared to Signal or SimpleX. The core philosophy around secure messaging here is that it can work in a way that avoids installation and registration by enabling users to host their own data. The proje
Anthropic’s recent security-incident document contains a bit about how CAPTCHAs are still frustrating Claude. In the transcript, the Claude model that is so powerful that Anthropic is gatekeeping access to it appeared to slam its virtual head against the wall solving a simple image identification test. In a test where the agent was asked to identify a shape that didn’t match the others displayed,
Training also requires content. The early norm was often: if material was accessible online, it was available for ingestion. Whether that is lawful remains contested across jurisdictions. Whether it was fair to creators is a separate question. Now many actors repeatedly crawl the same web. This imposes costs on the people and institutions that produced the value. Wikimedia reported that bots gener
My conclusion is blunt: the unchecked race of ever larger, privately controlled training bets has to stop. I do not mean research, every new model, or the use of LLMs. I mean that this cannot continue as normal. OpenAI and Anthropic appear to recognize at least part of the problem. Some industry leaders have called for pacing the frontier; OpenAI says it would slow or stop when risks cannot be suf
I can now turn this analysis into personal conclusions. The first level is political. Binding regulation of frontier-model training must come. Not because every new model is illegitimate, but because the present race allows individual actors to commit energy, water, hardware, data, and financial resources while shifting part of the risk onto everyone else. Voluntary promises cannot resolve an ince
Policy and procurement should prefer the smallest model that can reliably perform a task. That is not a ban on general-purpose or frontier models. Some problems may justify them. But “larger” should not be accepted as a synonym for “better.” Not every nail requires an orbital hammer. Smaller, specialized models can reduce computation, memory, hardware, and energy requirements in both deployment an
My wife and I are better at creating conspiracy theories than any of these MAGA assholes. RFK Jr. is now claiming a possible link between autism and chemtrails. Even if it were possible to have a linked cause, think about the general logistics involved in making such a thing happen. A better conspiracy is this. BTW, this is all sarcasm. Pharmaceutical companies are adding autism causing chemicals
Good morning from Minneapolis, where we're all fruit salad. All in the same bowl. Having clawed our way to Friday, it's 61℉; light rain; light winds; Fair barometer; visibility 10 miles at the airport. - Widespread steady soaking rain through today. - Heavy rain and localized flooding threat greatest near the Iowa border this evening. - Seasonable temperatures with mainly dry conditions expected f
The European Commission said on Sept. 18 that it would not propose an EU-wide windfall tax on energy companies, despite a joint request from Germany, Portugal, Austria, Spain, Italy and Poland. The six governments want a common levy on extraordinary profits linked to the energy-price shock caused by the war in Iran, but Economy Commissioner Valdis Dombrovskis said member states can introduce such
German minister says companies are ‘exploiting situation’ in Middle East, as sky-high prices become major domestic issue for leaders European governments have discussed imposing a bloc-wide windfall tax on energy companies, as near-record fuel and gas prices pile pressure on leaders desperate to contain mounting public discontent and the challenge of the far right. With elections due next year in
🚨🇰🇷 LOOKPIN dataset allegedly exposing 2.7M+ customers and 11.5M+ orders offered for sale ⠀ LOOKPIN is a South Korean men's fashion platform and mobile app offering clothing discovery and online shopping services. ⠀ A forum actor using the handle "seraphims" claims to have breached LOOKPIN in September 2026 using an authentication bypass involving SQL injection, allegedly exfiltrating customer,
Is there any “wiki” style project where one can refer to past events and disclosures about leaders, companies and celebrities after all “this blows over” in “a few years”? I wanna know when I apply for a job in 2030 what the CEO said in 2021/22 before laying off all their staff. (Eg, Mullenwig, Zuckerberg suddenly pivoting back to sounding woke again) I wanna know which Billionaire strong armed ot
I'm trying to get # FediHired so I made my first post on LinkedIn: "As part of my quest to always stretch myself I have been coding at home with nothing but my IDE, my brain, and reference material. I recently implemented, in Kotlin, an event-driven state machine with a limited size queue, and a pull consumer that consumes arbitrary sequences of events based on composable predicates. Because it's
I'll be offering an Introduction to Linux Kernel Exploitation training in English, to be held online next year (likely September). This will be a whole-month training, totaling 48 hours split across 12 sessions. We will have live classes with me, and the sessions will be recorded and provided to students within 24 hours for those who can't join live. The classes will happen on Monday, Wednesday, a
This summer brought plenty of apples in our three kinds of apple trees. I had the luxury of tasting them all and deciding on the mix to have juiced. And nailed it! Last time I used mostly the sweetest ones and the cider had very little apple taste in it. This one is very much apple tasting. Primary fermentation went on for over three weeks, and the result is bone dry, just like my type 1 diabetic
Just got laid off and I thought just in case I find something new before the notice period ends I should probably take all my remaining vacation days for the year, which turned out to be an entire month (plus two weeks that I've already planned for Christmas). I'm planning to go to a friend's lakeside shack in the woods for one or two weeks, then I gotta come back because friends are visiting from
🤓 Today at unprompted.au I presented some of the experimentation I did with the new model Jev from TypeSafe AI. I used Jev to evaluate AI agent drift and misalignment to detect if my agent is going rogue or being compromised. Jev is a model used for data classification and decision-making. It uses 3 different primitives: - Noul: Is the agent off-task? - Score: How severe is the drift? - Choice: W
Atlassian Patches 161 Vulnerabilities Across Data Center and Server Products Atlassian's September 2026 security bulletin addresses 161 vulnerabilities, including 17 critical flaws, across its Data Center and Server products like Confluence, Jira, and Bamboo. These vulnerabilities allow for remote code execution, man-in-the-middle attacks, and unauthorized session takeovers. **If you run Atlassian
everyone says "form a union" or "quit your job". weak aura. the real move is psychological warfare. yesterday my manager asked me why I missed the deadline. I simply looked at him, blinked slowly, and said "what deadline? you've been in a coma for 3 years, Richard. please wake up." he started crying. I took his wallet. I've been studying applied workplace sociopathy and mind control tricks from
Chatto is a small and fast chat server that looks like other chat products as a service, but can be self-hosted. It is privacy first, uses very little resources and written in Go and Svelte, is reasonably fast. It does have a working Bot API, and I tested that today with https:// github.com/isotopp/chatto-thre addump . That is not a bot, but a command line program using the Bot API. It will take a
An idea for a "reaction" feature for managing responses in a thread... based on vectors. I'm thinking of two vectors: one for time and one for logical agreement. And use fuzzy maths. The fuzzy part comes in by having a value somewhere in the range [0,1], with the midpoint 0.5 representing a "neutral" stance. The timewise vector would be how important you rate it to spend time on the tagged toot (w
The European Commission said Italy may have some room to manage liquidity from savings on loans funding its National Recovery and Resilience Plan, potentially supporting a planned 2027 vehicle-tax exemption for small cars and motorcycles. Economy Commissioner Valdis Dombrovskis made the comments after an informal Ecofin meeting in Dublin, responding to statements by Italian Economy Minister Gianca
🔐 SimpleX Chat launches privacy-preserving supporter badges with larger file limits SimpleX Chat has introduced supporter badges in the v7.1 beta, giving users additional features without tying the purchase to a persistent account or identity. ⠀ Supporter badge perks include: • Files up to 2 GB instead of 1 GB • File storage extended to 7 days • Badge displayed on your SimpleX profile ⠀ A higher
"Nie znam na tyle dobrze obecnej polskiej polityki partyjnej, żeby udzielać polskiej lewicy rad. Ale na podstawie naszych doświadczeń powiedziałabym: kluczowe jest poważne traktowanie konkretnych potrzeb ludzi." 🗳️🇩🇪 krytykapolityczna.pl/swiat/niemie... Podczas gdy polska lewica od lat walczy o odzyskanie wpływu na społeczeństwo, Die Linke przeżywa niezwykły renesans, choć jeszcze niedawno była
For the first time ever, a self-flying airplane has flown 3,199 miles to complete a journey across the US. A single-engine Cessna Caravan, managed by aerospace firm Joby Aviation, completed the transcontinental voyage without a human pilot onboard, relying entirely on autonomous computing architecture to navigate every stage from runway taxiing to takeoffs and touchdowns. The robotic aircraft depa
The issue of # Meta 's # PervertGlasses is not just about perverts obviously, and not even about "filming in public". It's about what happens to those pictures and videos. I never used to care about being caught in a private photo, because it used to be extremely unlikely to go anywhere but a family album book, and probably never seen again. Now I know the picture and lots of metadata (including d
Architecture is not only the structure of the system you have. It is the structure that determines how safely the system can become something else. Writing another chapter of Guerilla Architecture today...summarizing my approach to software these days. My unit of analysis isn’t the class or object relationship. It’s the flow of information through a changing sociotechnical system: code, data, infr
VULNARCHIVE and GNA 1988: Automated Vulnerability Identifier Allocation in a Federated GCVE Ecosystem. One of the core ideas behind GCVE is that vulnerability identification does not need to depend on a single central authority. Independent GCVE Numbering Authorities (GNAs) can operate their own processes, allocation policies and publication infrastructure while remaining interoperable with the wi
In August I wrote about the anchor that survives every cloud migration: the login. Move every workload to a European cloud and you can still reach it all only through an American identity provider. Crown jewels moved, keys left with the old landlord. The detail that made the piece worth writing is in Microsoft's own documentation. Azure Local's fully disconnected tier does not support Entra ID. Mi
🚨🇫🇷 Hôpital Paris Saint-Joseph dataset allegedly exposing 15K+ records Hôpital Paris Saint-Joseph is a nonprofit private hospital in Paris providing care across multiple medical specialties. ⠀ A forum actor using the handle "weykofa" claims to have obtained and released data associated with Hôpital Paris Saint-Joseph, advertising 15,061 individual records and information tied to 738 doctors. ⠀
[Sentinel/US-West] # opencanary analysis for yesterday Summary: 🔄 Total Connection Attempts: 51145 👤 Unique Usernames: 524 🔑 Distinct Passwords: 1248 🌐 Unique Attacker IPs: 853 Port Popularity (Port / Count): 🖥️ RDP: 27251 🖥️ VNC: 11651 🔐 SSH: 9709 🗄️ SMB: 1268 💻 Telnet: 1114 🗃️ MSSQL: 69 🔴 REDIS: 51 🐬 MySQL: 29 📂 FTP: 2 🐙 GIT: 1 Top 10 Usernames (Username / Count): 👤 anonymous: 136
TO THOSE WHO WERE HARRASSED A document on why the UTTP archive exists, and who it was built for ================================================== FOREWORD This document is not about UTTP. It is about the people who were on the receiving end of UTTP and never got an answer. If you were raided, doxed, threatened, extorted, mass-reported, or harassed by accounts that called themselves UTTP - this is
>Google's September 2026 Pixel Update Bulletin contains patches beyond what's in that month's regular Android Security Bulletin. According to GrapheneOS, some of those extra patches touch standard Android platform code, the kind that runs on non-Pixel devices, not just Pixel-branded hardware. > >None of that platform-level code has reached the regular monthly bulletin or the private preview patche
I'm kinda torn about Special Ops Lioness so far... On the one hand yeah obviously it's pro-state* copaganda porn... On the other hand it does spend significant time highlighting how chauvinistic bureaucracy cost so much in terms of human lives, both sociopolitically and mental-psychologically, both at the individual level and at the community/nation-state level... War is evil and a failure of dipl
arXiv:2609.21088v1 Announce Type: new Abstract: Indirect prompt injection (IPI) remains a central safety and security challenge for large language model (LLM) systems because standard transformers lack architectural notion of source authority. Retrieved documents, user inputs, and system instructions are all processed through the same undifferentiated attention mechanism, forcing the model to infe
arXiv:2609.21103v1 Announce Type: new Abstract: Modern networks are large in scale and heterogeneous in configuration, making manual policy management increasingly impractical. Intent-Based Networking (IBN) addresses this by automating the translation of high-level operator goals into low-level network configurations. Yet existing IBN systems rely on static heuristics and fixed-feature classifiers
arXiv:2609.21217v1 Announce Type: new Abstract: Automotive Ethernet carries heterogeneous multi-protocol traffic in modern in-vehicle networks, where labeled attack data are rarely available and the strongest prior unsupervised detector still relies on handcrafted traffic features. This article presents X-SPUR, an explainable, surprisal-based, protocol-aware unsupervised reasoning framework that i
arXiv:2609.21340v1 Announce Type: new Abstract: Empirical identity leakage from released text is increasingly driven by attackers that combine large language models (LLMs) with auxiliary knowledge to link documents to individuals. Existing audits typically report success rates for specific attack pipelines but lack finite-sample statistical guarantees, while training-time protections such as diffe
arXiv:2609.21419v1 Announce Type: new Abstract: Software Bills of Materials (SBOMs) have emerged as a key mechanism for software supply chain governance in enterprise architectures. However, their adoption across organizations remains limited due to concerns about exposing sensitive dependency information. To address this limitation, we propose TrustBOM, a scalable architecture for confidentiality
arXiv:2609.21426v1 Announce Type: new Abstract: Ransomware increasingly fragments its file operations across temporary and intermediate files, scattering the semantic context that links individual I/O events to an overarching encryption campaign. This fragmentation defeats existing detectors that reason over isolated file streams -- whether pattern-based methods that match rigid event sequences or
arXiv:2609.21484v1 Announce Type: new Abstract: Homomorphic encryption (HE) has emerged as a promising approach to privacy-preserving machine learning (PPML), enabling computation directly over encrypted data. In HE-based PPML, a client submits an encrypted input to the server, which evaluates models such as large language models (LLMs) without access to the underlying plaintext. However, we ident
arXiv:2609.21515v1 Announce Type: new Abstract: Third-party adapters for open-weight language models ship as opaque weight matrices; a recipient cannot check whether an adapter hides a backdoor without trusting the publisher or inspecting the weights, the publisher's core asset. For one important class (payloads placed where a safety monitor is structurally blind), detection is unsound as a defens
arXiv:2609.21532v1 Announce Type: new Abstract: In cryptography, critical sets of Latin squares have particularly been implemented to design secret sharing schemes. A main problem in these cryptographic protocols arises from absent holders of pieces of information that are common to different critical sets, because they become indispensable to recover the secret. This paper solves this problem by
arXiv:2609.21569v1 Announce Type: new Abstract: Recent generations of Apple MacBooks embed an inertial measurement unit (IMU) within their unibody chassis for device orientation and motion sensing. However, this IMU inadvertently captures not only intended device-level information but also subtle physical vibrations from user interactions and the surrounding environment. These signals establish a
arXiv:2609.21573v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) improves large language models by grounding outputs in external knowledge sources, but this dependency also creates a surface for poisoning attacks. This paper introduces Micro-Collaborative Poisoning, a distributed attack in which a false target claim is divided across multiple locally plausible documents instead
arXiv:2609.21686v1 Announce Type: new Abstract: Privacy leakage in LLM agents is commonly evaluated within individual components such as memory, retrieval, or tool-use pipelines, which makes it difficult to distinguish internal exposure from information that an external observer can actually recover. We present CIPL (Channel Inversion for Privacy Leakage), a channel-aware evaluation framework for
arXiv:2609.21713v1 Announce Type: new Abstract: Edge AI accelerators are increasingly deployed in safety-critical environments, where model outputs may control physical actuators, make access-control decisions, or trigger alarms. In these settings, runtime failures often remain undetected because model corruption, distribution shift, and adversarial inputs can still produce well-formed, confident
arXiv:2609.21728v1 Announce Type: new Abstract: We present an architecture that enables data owners to combine private data into data pools using Trusted Execution Environments (TEEs) and manage these pools by issuing narrowly scoped computational rights, encoded as Digital Rights Tokens (DRTs), to third-party data analysts. Each DRT binds specific open-source code to a pool and is issued and rede
arXiv:2609.21793v1 Announce Type: new Abstract: Defenses against jailbreak attacks on Large Language Models (LLMs) operate at different pipeline stages, such as input modification or output guard, but it remains unclear which defenses to deploy at each stage and how to combine them. Prior empirical studies, fragmented by inconsistent attack-success-rate definitions and experimental settings, have
arXiv:2609.21858v1 Announce Type: new Abstract: Large language models (LLMs) have achieved state-of-the-art performance across a wide range of tasks, motivating two important aspects of deployment: inference efficiency and output provenance, which can be tackled by speculative sampling and watermarking, respectively. However, recent works have shown that combining these two goals is highly nontriv
arXiv:2609.21957v1 Announce Type: new Abstract: A companion paper found that what a defender must block over time has units: bits per period [Takashita, 2026a]. This paper sets it. Hold every crossing in escrow for one period, audit each held item independently with probability r, and revoke the window if any audit catches something. An adversary crossing k times, each carrying c bits, expects kc(
arXiv:2609.22018v1 Announce Type: new Abstract: Given a supersingular elliptic curve $E/\mathbb{F}_{p^2}$, the $\mathsf{OneEnd}$ problem asks for a non-scalar endomorphism of $E$. By known reductions, solving this problem also solves the supersingular endomorphism ring and isogeny problems. Wesolowski obtained exponent $1/3$ under an assumption on the factorization of a small degree, whereas the p
arXiv:2609.22076v1 Announce Type: new Abstract: APort Vault is a benchmark for payment authorization in tool-using AI agents. It replays 4,371 attacks written by humans against a live payment agent during a public capture-the-flag event, across 14 models from 8 labs, five policy configurations and two replay tracks, with and without a deterministic pre-action check implementing the Open Agent Pass
arXiv:2412.07151v1 Announce Type: cross Abstract: Distributed model training needs to be adapted to challenges such as the straggler effect and Byzantine attacks. When coordinating the training process with multiple computing nodes, ensuring timely and reliable gradient aggregation amidst network and system malfunctions is essential. To tackle these issues, we propose \textit{dSTAR}, a lightweight
arXiv:2609.20982v1 Announce Type: cross Abstract: Reinforcement learning (RL) controllers have been recently adopted for Unmanned Aerial Vehicles (UAV) navigation and control. However, they are susceptible to action-space attacks that overwrite the action commands after the policy generates them and before the actuators execute them. While most existing defenses target attacks on the policy's inpu
arXiv:2609.21284v1 Announce Type: cross Abstract: Long-running AI agents outlive initiating processes through credentials, delegated tasks, queues, callbacks, reservations, and provider-side operations. Cancellation, process exit, and credential revocation neither close every pre-cut carrier nor distinguish independently authorized shared work. We define root-scoped authorization quiescence: for e
arXiv:2505.06304v5 Announce Type: replace Abstract: The protection of Intellectual Property (IP) for Large Language Models (LLMs) has become a critical concern as model theft and unauthorized commercialization escalate. While adversarial fingerprinting offers a promising black-box solution for ownership verification, existing methods suffer from significant limitations: they are fragile against do
arXiv:2510.07479v2 Announce Type: replace Abstract: We present $\mathsf{Miranda}$, the first family of full-domain-hash signatures based on matrix codes. This signature scheme fulfils the paradigm of Gentry, Peikert and Vaikuntanathan ($\mathsf{GPV}$), which gives strong security guarantees. Our trapdoor is very simple and generic: if we propose it with matrix codes, it can actually be instantiate
arXiv:2511.14301v4 Announce Type: replace Abstract: Transformer-based models are highly susceptible to backdoor attacks via supervised fine-tuning (SFT). To red-team existing data-poisoning defenses, prior work has increasingly focused on stylized triggers, synthetic artifacts, and token-level perturbations designed to evade detection. However, this trend has shifted threat models away from natura
arXiv:2607.01356v2 Announce Type: replace Abstract: Cyber-physical systems (CPSs) can be compromised through memory corruption vulnerabilities, which can result in safety violations. Existing techniques mostly focus on detecting such attacks but respond by terminating or halting execution upon attack detection, which is not acceptable in CPSs as interrupted tasks can have catastrophic consequences
arXiv:2502.02542v5 Announce Type: replace-cross Abstract: A reasoning language model (RLM) generates costly reasoning tokens, often hidden from the users, that help it excel at many tasks. Our Overthink attack targets RLM-based applications (such as chatbots or coding agents) that rely on external context by forcing these models to generate substantially more reasoning tokens while still producing
arXiv:2609.21232v1 Announce Type: new Abstract: A substantial body of empirical evidence suggests that stochastic mortality models ignoring long range dependence tend to underestimate life expectancy, which may lead to profound implications for pension schemes and funding arrangements. This paper addresses the modelling of stochastic mortality via a mixture of a fractional L\'evy process and stand
arXiv:2609.21271v1 Announce Type: new Abstract: Nested clustered optimization allocates within each cluster from the cluster's own covariance block and then across the resulting cluster portfolios. Block inversion says the unconstrained minimum-variance portfolio has the same two-tier shape, with each block replaced by its Schur complement against every other asset. Conditioning instead on one kno
arXiv:2609.21291v1 Announce Type: new Abstract: We develop an efficient Monte Carlo simulation scheme for pricing options under the Ornstein-Uhlenbeck driven stochastic volatility model via the operator splitting approach. With an ingenious splitting of the governing stochastic differential equations, our operator splitting scheme admits analytic solutions in all sub-steps, so its implementation i
arXiv:2609.21301v1 Announce Type: new Abstract: The standard Euler discretization schemes for numerical option pricing under stochastic volatility models are known to exhibit high biases and potential unreliability. The alternative use of the exact (unbiased) simulation approach invariably involves numerical evaluation of integrated variance (and / or volatility) conditional on terminal variance (
arXiv:2609.21684v1 Announce Type: new Abstract: We study a representative-agent Epstein-Zin economy with geometric dividends and a hidden finite-state Markov drift. We allow the price-dividend ratio to contain an additional positive, absolutely continuous valuation factor and, within the class $\mathfrak C$ defined below and under the regularity, admissibility, and positivity conditions of our mai
arXiv:2609.22052v1 Announce Type: new Abstract: Catastrophe (cat) bonds overcome some lack of reinsurance by sourcing capacity from the wider capital markets. We present a new type of cat bond addressing the known trade-off between moral hazard and basis risk. As our main contributions we propose a trigger mechanism which is entirely transparent and simpler to evaluate compared to indemnity modeli
arXiv:2609.21688v1 Announce Type: cross Abstract: We study random signed Takagi--Landsberg bridges with independent Rademacher Faber--Schauder coefficients. At index $H=1/2$, we prove that every fixed deterministic refining sequence of partitions with vanishing mesh yields quadratic variation $t$, almost surely and uniformly in $t\in[0,1]$. For partitions that need not be refining, the mesh condit
arXiv:2606.08228v2 Announce Type: replace Abstract: Filter-gated algorithmic trading systems on decentralised exchanges reject most candidate tokens they evaluate, yet the observed forward market trajectory of those rejected candidates is rarely measured on the same live venue that produced the rejection. This paper introduces Post-Rejection Follow-up Sampling (PRFS), an observational measurement
arXiv:2607.10542v2 Announce Type: replace Abstract: We study an optimal portfolio and consumption problem in a regime-switching multi-name credit market with default contagion. Default events not only generate direct portfolio losses but also alter the default intensities of surviving securities. Under a Cobb-Douglas utility, the homogeneity property reduces the associated Hamilton-Jacobi-Bellman
arXiv:2511.21646v2 Announce Type: replace-cross Abstract: Optimal control of interacting particles governed by stochastic evolution equations in Hilbert spaces is an open area of research. Such systems naturally arise in formulations where each particle is modeled by stochastic partial differential equations, path-dependent stochastic differential equations (such as stochastic delay differential e
Media workers are grappling with increasingly volatile press climate in US, with some detained and even deported Banished: Trump’s secret deportation deals What is Trump’s third-country deportation policy? As hundreds of masked and armed US federal agents besieged the streets of Minneapolis, turning the midwestern city into the frontline of the Trump administration’s unprecedented crackdown on imm
Managing WinSCP in the enterprise? If # WinSCP is deployed across your environment, how are you managing its configuration and security controls today? WinSCP supports a number of registry-based administrative settings that can be used to enforce behaviour across the application, including controls around credential storage and preventing passwords from being stored in plaintext. I decided to turn
Ukraine Brief — 2026-09-21 Ukraine's first Pelican ballistic missile strike on Moscow and Operation Vivaldi's ground gains expand Kyiv's multi-domain capability, but Russian industrial targeting has eliminated 90% of national steel production while a four-week Rada recess forecloses legislative response to mounting fiscal pressure. Russia will very likely retaliate against Ukrainian energy or indu
'[Aircraft bits] are primarily produced in the United States, France, Germany, and the United Kingdom … building high-performance turbofan engines or advanced flight-control systems remains virtually impossible without Western technology. '... a legal and regulatory framework that dictates global airspace access. At the top sits the [ICAO] … the operational teeth of this system [Chicago Convention
What GAO Found The Federal Aviation Administration (FAA) has identified electromagnetic spectrum-related threats, including spoofing and jamming, to the National Airspace System (NAS) and international flight routes. However, FAA has not completed risk and mitigation assessments, and updated security documentation needed to address these threats. Additionally, FAA did not have a defined, real-time
What GAO Found Many countries, including the U.S., are experiencing an increase in both the number and the proportion of older adults in their populations. According to the World Health Organization, the share of the global population aged 60 and over is expected to more than double from one billion in 2020 to 2.1 billion by 2050. As countries experiencing population aging take steps to address th
This is the Wikipedia definition for human intelligence. Artificial intelligence (in the form of LLMs) does literally none of these things. It doesn't act by itself, it isn't self-aware[1], it doesn't learn[2], it doesn't understand or form concepts[3], it doesn't logic or reason[4], it doesn't recognize patterns[5], plan, innovate, solve problems, make decisions[6], retain information or even use
~ Personal news ~ I left TechCrunch a few days ago. First, I will now focus on finishing my book about Hacking Team and the history of government spyware. After that, and in the meantime as well, I will be freelancing. So if you are looking for some cool stories about cybersecurity and surveillance; or help researching and reporting for a larger project, and similar services, please let me know. Y
Security updates have been issued by AlmaLinux (kernel, perl-Net-DNS, sudo, tomcat, and tomcat9), Debian (chromium, gimp, libde265, libevent, linux-6.12, ruby-jwt, and unbound), Fedora (asterisk, chromium, doctl, dovecot, evolution, firefox, forgejo, freeciv, freeipa, gegl04, gimp, libheif, nss, opkssh, parted, ruby, stb, thunderbird, unbound, and webkitgtk), Mageia (bind, gawk, gdk-pixbuf2.0, gra
Imagine working with an immigration lawyer to prepare an application several hundred pages long, then the decision comes back in under a minute, with the officer looking at several files at once. How would you feel if you were the applicant? Immigration, Refugees and Citizenship Canada measures how long it takes officers to process visa and immigration applications. Should be a win, right? Not so
At, GPTree, we run several coding agents across our team on one repo using parallel worktrees. Apart from wasted time reviewing and fixing conflicts at PR time, the failures that hurt the most are when multiple plans or tickets cause architecture changes that cannot both be true. Ex. one agent replaces a class while another one is in the process of extending it. Git only notices if the resulting p
One server. Ten simultaneous phishing campaigns. Nigerian-origin actor confirmed. Matrix flagged zoom4usinvite[.]space as an open-directory staging server. The server backup left world-readable the day before the crawl revealed the full picture: alongside Zoom and Adobe/ClickFix droppers, the actor runs real-time Adversary-in-the-Middle kits for Google (2SV bypass), Microsoft (Authenticator + SMS
This bold ad on the inside back cover of the August Atlantic magazine caught my eye. The two strong sentences rang bells in my head, but it only made sense after I went to the TearThePaperCeiling.Org. As a high school dropout without a bachelor's degree, everything this organization is saying made clear sense to me. College is not for everyone, and not everyone should be forced to go. It's expensi
🚨 GoreSpiders recruits ransomware distributors with a claimed 25% share of victim payments A forum actor using the handle "GoreSpiders" is advertising a ransomware distribution partnership on PwnForums, offering participants a share of ransom payments from victims they bring in. ⠀ The actor claims distributors receive 25% of each payment, with the share potentially increasing to 40% or more for t
Need to sell my beloved deepkin army as my daughter is off to Uni so need all the money i can get to help support ebay.io/m/YQlZef Idoneth Deepkin army painted to a very high standard for Warhammer age of Sigmar. White armour is in fact pearlescent which doesn’t show well in pictures. GW flight stands replaced with custom stands from magnet baron on all applicable models Volturnos Ahkelian King Lo
TIL: oggi ho scoperto la “regola del buon vicino” di Aby Warburg applicata dal grandissimo Umberto Eco alla sua amata biblioteca. In pratica: il libro che stai cercando è utile, ma quello accanto potrebbe esserlo ancora di più per affinità. Quasi quasi la applico anche ai miei vinili. Serendipità, con scaffali. ⬇️ "Ricostruire la mitica biblioteca di Umberto Eco Decine di migliaia dei libri che co
Ingles Markets is proud to work with local food and beverage makers and farmers. You’ll find local and regional products throughout our stores from in-season fruits and vegetables in our Produce Department, artisan breads from City Bakery and Annie’s Breads in the bakery, local meat and sausage from Hickory Nut Gap Farms, trout from Sunburst Trout Farm, snacks from Poppy Handcrafted Popcorn and As
🚨🇺🇸 Cyberattack disrupts internet service across Eagle Mountain, Utah Direct Communications, an internet provider serving much of Eagle Mountain, says unusually heavy traffic consistent with a distributed denial-of-service (DDoS) attack caused repeated outages. ⠀ Reported disruption includes: • Slow speeds, packet loss and complete internet outages • Connectivity problems involving upstream pro
Follow the day’s news live. Get our breaking news email , free app or daily news podcast Jim Chalmers says migration is necessary to support Australia’s ageing population, and can help the nation as birth rates tumble. The three-yearly intergenerational report found that the declining birth rate meant more Australians will be dying than being born in 40 years’ time – the first time the report had
arXiv:2609.22510v1 Announce Type: new Abstract: As LLM applications integrate with external tools, they are increasingly exposed to indirect prompt injection (IPI), where adversarial instructions are embedded in retrieved content. Conventional IPIs fire on contact: the moment an agent ingests the content, it carries out the instruction. We introduce the explosive prompt, a conditional payload that
arXiv:2609.22525v1 Announce Type: new Abstract: The Precision Time Protocol (PTP) provides the time and phase synchronization required by disaggregated Open Radio Access Networks (O-RAN). Yet, in current open fronthaul deployments, PTP traffic lacks mandatory authentication and integrity protection, leaving synchronization vulnerable to spoofing, replay, and delay manipulation attacks that can deg
arXiv:2609.22573v1 Announce Type: new Abstract: LLM agents translate natural-language context, which may include attacker-controlled text, into privileged tool calls, so authorization must remain effective even when an agent is prompt-injected or adversarially steered. The Model Context Protocol (MCP) has become a widely adopted interface for this boundary, yet its official SDKs' authentication an
arXiv:2609.22664v1 Announce Type: new Abstract: Research on large language model agents for penetration testing is evaluated almost entirely by capability: whether the agent captures a flag or reproduces a proof of concept. That metric suits a benchmark but is silent on the properties that decide whether an autonomous agent can be used in an authorized engagement: whether a reported finding is tru
arXiv:2609.22724v1 Announce Type: new Abstract: Mobile agents powered by foundation models now automate complex, multi-step workflows on real devices, but their trajectories can violate app-specific security policies. Existing trajectory-level defenses rely on LLM prompting or rigid rules, and thus fail to support fine-grained, natural-language policies that generalize across apps and tasks. In th
arXiv:2609.22792v1 Announce Type: new Abstract: LLM agents are increasingly used for security tasks: vulnerability discovery, exploit reproduction, and patch generation. Improving them at the model level demands expert demonstrations or computable rewards, which security tasks rarely offer: traces are costly, failures hard to diagnose, rewards sparse, and non-computable. Efforts thus shift to the
arXiv:2609.22818v1 Announce Type: new Abstract: Memory-poisoning defenses for LLM agents are typically evaluated by their ability to prevent attacks. However, the traffic they process is rarely adversarial. The cost of implementing a defense is paid with each interaction, while its benefits are only seen in a small percentage of cases. We developed a measurement setup that keeps the memory backend
arXiv:2609.22835v1 Announce Type: new Abstract: Machine learning methods for Android malware family classification have achieved high accuracy, but their application is hindered by two major challenges. First, the widely used code obfuscation severely disrupts the automated labeling process and introduces substantial label noise into training datasets. Second, training datasets often exhibit sever
arXiv:2609.22900v1 Announce Type: new Abstract: A SUPL\_INIT message is a network-initiated trigger that can be sent to a handset using an SMS to unilaterally start a location session: on receipt, the handset is instructed to determine its own position and report it, together with an identifier such as its IMSI, to a server specified in the message, without any action by the phone's user. The conc
arXiv:2609.22949v1 Announce Type: new Abstract: Existing prompt injection research focuses on single-model chatbot scenarios, where an attacker manipulates one LLM through crafted input. Multi-agent systems amplify this threat through three mechanisms absent from single-model settings: inter-agent message passing creates injection channels invisible to perimeter defenses, shared tool access enable
arXiv:2609.22961v1 Announce Type: new Abstract: Agentic systems increasingly invoke tools, services, data, and other agents across organizational boundaries, yet a relying party cannot assess a delegated action solely from producing-domain controls and records. This paper develops Trustworthiness as a Service (TaaS) through a synthesis of trustworthy-AI governance, agent security, distributed trus
arXiv:2609.22981v1 Announce Type: new Abstract: We present a dual-locking method for securing trained neural networks that combines key-driven index permutation with PIN-based watermarking based on Sparse Quantization Index Modulation (QIM). Cryptographic randomness is introduced by independently applying a uniform random permutation to each row of adaptively selected index vectors. A robust blind
arXiv:2609.23042v1 Announce Type: new Abstract: Pre-deployment secret scanning operates only on source code, never on what a production application serves. We document two exploitation chains in which Azure AD client credentials and APIM subscription keys from production JavaScript bundles enabled account takeover and mass data exposure. An authorized engagement covered approximately 2,000 enterpr
arXiv:2609.23097v1 Announce Type: new Abstract: Internet of Things (IoT) systems are increasingly deployed in smart homes, transportation, energy systems, and critical infrastructure. This broad connectivity improves service intelligence, but also enlarges the attack surface of IoT networks. Machine Learning (ML)-based Intrusion Detection Systems (IDSs) are widely used to identify malicious networ
arXiv:2609.23173v1 Announce Type: new Abstract: Hardware trojan (HT) attacks against CPUs typically assume threat scenarios where an attacker targeting a system with a trojanized CPU is able to execute arbitrary code (i.e. machine-level instructions) to reliably interact with the implanted trojan. On end-user devices (i.e., mobiles, laptops), achieving arbitrary code execution in practice requires
arXiv:2609.23193v1 Announce Type: new Abstract: As Large Language Model (LLM) APIs become increasingly integrated into privacy-sensitive workflows, ensuring inference-time privacy without compromising task utility remains a major challenge. Existing approaches preserve most of the original semantic content to maintain downstream performance, but this also leaves exploitable cues for reconstructing
arXiv:2609.23218v1 Announce Type: new Abstract: Automated kernel bug discovery has advanced rapidly. Continuous fuzzing and static analysis systems, such as syzbot, now expose Linux kernel bugs at a scale that downstream processes struggle to absorb. Yet a crash report is only the beginning. Before a bug is eliminated, it must be triaged, understood, patched, validated, reviewed, integrated, and o
arXiv:2609.23396v1 Announce Type: new Abstract: We propose a stealth framework that eliminates detection of hooking tools such as Frida and Xposed in secured mobile environments by replacing static configurations with dynamic evasion tactics. In contrast to existing approaches that apply these techniques independently, the framework introduces a unified runtime control layer that systematically co
arXiv:2609.23405v1 Announce Type: new Abstract: Address-Poisoning Transfer (APT) is a prevalent blockchain phishing scam in which a scammer poisons a victim's address book by generating a transfer with a phishing address that looks similar to a benign address that the victim has previously interacted with. Although simple, APT phishing attacks have cost users millions of dollars in recent years, w
arXiv:2609.23498v1 Announce Type: new Abstract: Agentic systems increasingly fulfill user requests through multi-step tool workflows over files, services, and external resources. In these workflows, isolated per-call checks can miss a workflow-level failure: each call may be locally admissible, but the sequence can exceed the authorization boundary established for the session. We identify this fai
arXiv:2609.23514v1 Announce Type: new Abstract: We propose endogenous interpretation: program, interpreter, machine, and derived execution language are not disjoint semantic objects but different parameterizations of one executable state-transition relation. Each realization carries an implicit constraint bias, the structural restrictions imposed by its instruction basis, state encoding, control t
arXiv:2609.23583v1 Announce Type: new Abstract: Firmware fuzzing is an effective technique for discovering vulnerabilities in embedded systems. However, existing coverage-guided firmware fuzzers typically obtain feedback through firmware instrumentation, hardware debug interfaces, or firmware rehosting, which requires access to the firmware source code or binary image. Such requirements are often
arXiv:2609.23766v1 Announce Type: new Abstract: Root cause analysis at a remote site is slow: evidence is scattered across pod logs, Kubernetes events and cluster-level objects, and many operators cannot send production logs to a hosted model at all. On-premise inference removes the second constraint but raises a question live-cluster benchmarks have not addressed: when one workstation GPU fixes b
arXiv:2609.23827v1 Announce Type: new Abstract: Current privacy-preserving mechanisms (PPMs) in Complex Event Processing (CEP) systems are unnecessarily restrictive, reducing the utility of data received by data consumers. This article presents a novel approach to preserve privacy in CEP systems, improving the utility of detected event patterns by dynamically adapting the noise added to an unprote
arXiv:2609.23889v1 Announce Type: new Abstract: Automated kernel vulnerability reproduction is essential for bug triage, patch validation, and regression testing, but still lacks an effective and efficient solution. The core challenge is twofold: a reproducer must first recover the trigger scaffold needed to reach the vulnerable state and determine the precise concrete values that actually trigger
arXiv:2609.23980v1 Announce Type: new Abstract: AI agents now report vulnerabilities faster than maintainers can review them. Reports often depend on security properties specific to the application, and require considerable human labor to process. To mitigate this, we introduce a framework for evaluating vulnerability reports via probes, executable checks of security properties. A reported exploit
arXiv:2609.24077v1 Announce Type: new Abstract: Privileged language-model agents can satisfy a new system task by displacing a healthy incumbent that depends on the same file, process, socket, lock, or capacity allocation. This failure arises because execution privilege determines whether an operation can run, not whether the requester may preempt the current resource owner. We present LeaseGuard,
arXiv:2609.24084v1 Announce Type: new Abstract: Open-weight large language models (LLMs) can be copied, modified, and redeployed behind black-box APIs, making post-release ownership verification difficult. Existing black-box fingerprints often rely on secret query-key pairs that reproduce predefined responses, and can therefore be easily disrupted by fine-tuning, pruning, quantization, model mergi
arXiv:2609.24280v1 Announce Type: new Abstract: Malware detection is a critical task in cybersecurity, and graph neural networks over control flow graphs have shown promising results for it. However, detectors are usually evaluated on a random split of a corpus collected over a single period, which cannot show how well a model generalizes to later samples. This study addresses that limitation with
arXiv:2609.24389v1 Announce Type: new Abstract: Deep learning-based malware detection has been widely adopted in security-critical services. Most detection methods rely on internal features extracted from APK files or runtime behavior. However, extracting these features is computationally expensive. This limits their use in large-scale, early-stage screening. Malicious apps may exhibit weak corres
arXiv:2609.24446v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly execute long-horizon workflows through external tools, allowing untrusted outputs to influence subsequent actions and exceed user authorization. Existing defenses isolate injected content or constrain execution with predefined plans and static policies, but these approaches are brittle under dynamic work
arXiv:2609.24550v1 Announce Type: new Abstract: The security of the modern web depends on the correctness of JavaScript (JS) engines, yet these complex systems remain vulnerable to high-impact bugs. A critical limitation of state-of-the-art fuzzers is the coverage plateau: once a fuzzer saturates the control-flow graph, edge coverage loses its ability to guide discovery. Because complex engine beh
arXiv:2609.24656v1 Announce Type: new Abstract: The fifth generation of mobile networks was standardised with an explicit mandate to close long-standing privacy and security gaps, mandating the concealment of the subscriber's permanent identity, resistance to generational downgrade, and protection against location tracking. Assessing whether these guarantees hold in operational networks, however,
arXiv:2609.24801v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed in production systems, raising concerns about their exposure to adversarial manipulation through prompt injection and jailbreak attacks. Classifier-based guardrails, such as Prompt Guard 2, are widely used as a first line of defense against such attacks, but their internal decision logic is large
arXiv:2609.24826v1 Announce Type: new Abstract: When a backdoor trigger activates the target response regardless of the triggered prompt context, the backdoor objective reveals itself. Challenging this trigger-sufficient formulation across the LLM backdoor literature, we introduce Opportunistic Backdoors (OPBackdoor), in which the backdoor objective is elicited only when the triggered prompt conte
arXiv:2609.22157v1 Announce Type: cross Abstract: KV-cache eviction methods decide which tokens to keep but not whether to evict at all, so a benchmark mean can hide a class of inputs on which compression drives accuracy from 99\% to 0\%. We reframe eviction as a per-input admission decision and show that inputs separate into a capacity-bound class, where eviction is catastrophic at every budget,
arXiv:2609.22284v1 Announce Type: cross Abstract: Talking-face (TF) deepfakes are detected unevenly by rPPG-based methods across generators. We study two lightweight visual-only cues, rPPG-derived waveforms extracted by RhythmFormer and lip-region discrete cosine transform (DCT) coefficients, on the seven TF methods of Celeb-DF++ under a subject-independent protocol. In-domain, lip-region DCT matc
arXiv:2609.22392v1 Announce Type: cross Abstract: Image steganography hides secret message within normal images, with most existing works relying on cover-preserving transmission. However, such a paradigm becomes vulnerable once the original cover is exposed or can be reliably approximated. In this paper, we propose StyleStegaNet, a stylized image hiding framework that replaces cover matching with
arXiv:2609.22590v1 Announce Type: cross Abstract: Soft errors are an increasing reliability concern for Deep Neural Network execution because they can corrupt parameters, leading to accuracy degradation. While conventional ECC offers strong fault protection, it incurs additional parity storage and computational overhead. Embedded-parity formats reduce storage cost by reusing the least-significant
arXiv:2609.22601v1 Announce Type: cross Abstract: The increasing reliance of autonomous AI agents on external and distributed knowledge sources introduces a fundamental challenge for decentralized information marketplaces: retrieval agents must evaluate the quality and relevance of data before purchase, while data providers must avoid revealing valuable information prior to guaranteed compensation
arXiv:2609.22968v1 Announce Type: cross Abstract: We study a one-player token game on an $N\times N$ board where tokens slide along diagonals or duplicate onto neighbouring ones to form a combinatorial rectangle $R\times S$. A conserved integer weight $W'$ and a strict monovariant guarantee $O(N^2)$-length solutions, placing the game in $\mathsf{NP}$. We prove that reaching a final position factor
arXiv:2609.23596v1 Announce Type: cross Abstract: With face-recognition models now embedded in everyday authentication and surveillance, recent works have pinpointed a critical weakness: these models remain acutely vulnerable to adversarial semantic edits. I.e., adversarially produced semantic alterations to the input, such as slight aging or pose changes, can induce misclassifications. Certain ex
arXiv:2609.23953v1 Announce Type: cross Abstract: AI agents that carry a multi-step computer task through on their own became ordinary tools in the past year, and the same autonomy is available to anyone whose task is harmful. We ask what that means for a relying party -- an insurer, a lender, an auditor -- whose evidence is a filed PDF. AgentForge-Bench measures how reliably an off-the-shelf codi
arXiv:2609.24036v1 Announce Type: cross Abstract: This paper presents an LLM-based system that translates natural-language access control policies (NLACPs) into executable Rego code for Open Policy Agent (OPA). It provides a modular, end-to-end pipeline for policy detection, component extraction, schema validation, linting, compilation, and automated test generation and execution. The system is de
arXiv:2609.24249v1 Announce Type: cross Abstract: Neural networks, both convolution or transformer based, are essential for modern computer vision systems. However, they are vulnerable to small perturbations, almost imperceptible to humans, which significantly alter the model's prediction. These adversarial attacks are often considered to be a significant threat to the implementation of neural net
arXiv:2609.24291v1 Announce Type: cross Abstract: Understanding the entanglement cost of non-local quantum computation (NLQC) is relevant to complexity theory, cryptography, quantum gravity, and related areas. A central special case is $f$-routing, motivated in part by quantum position verification. Proving lower bounds on its entanglement cost in the fully robust setting has been a major open pro
arXiv:2609.24359v1 Announce Type: cross Abstract: Image forensics is increasingly an open-world problem: manipulations range from fully synthetic images to localized edits, splicing and swapping, while most forensic detectors remain specialized to a single manipulation family. Agentic AI has recently emerged as a promising solution. In principle, such systems can assess the reliability of individu
arXiv:2609.24537v1 Announce Type: cross Abstract: Video recording on smart glasses exposes more than faces. Continuous capture reveals full-body biometric signatures, including gait, posture, and silhouette, that enable person re-identification (ReID) even after conventional face sanitization. We present MIRAGE, a three-tier architecture for privacy-preserving smart glasses that enforces full-body
arXiv:2609.24994v1 Announce Type: cross Abstract: As large language models (LLMs) are increasingly used to automate digital interactions, users can leverage LLM-generated text as cover for covert communication within seemingly benign conversations. Existing LLM steganography, however, is predominantly white-box, requiring the sender and receiver to share the cover statistics, typically through acc
arXiv:2507.02699v2 Announce Type: replace Abstract: Large Language Model (LLM) email agents have emerged as pivotal autonomous assistants, serving as a critical root of trust for digital identity by managing sensitive communications and authentication workflows. Despite their importance, the prompt injection (PI) resilience of the real-world LLM email agent ecosystem remains poorly understood. Exi
arXiv:2508.20517v2 Announce Type: replace Abstract: Cross-chain bridges enable asset and state transfers across heterogeneous blockchains, but their complex cross-domain interactions introduce new attack surfaces that are difficult to monitor using traditional single-chain analysis methods. Existing approaches often focus on isolated on-chain behaviors and fail to capture the multi-stage execution
arXiv:2509.08704v2 Announce Type: replace Abstract: In this paper, we study the problem of privacy audit in one run and show that our method achieves tight audit results for various differentially private protocols. This includes obtaining tight results for auditing $(\varepsilon,\delta)$-DP algorithms where all previous work fails to achieve in any parameter setups. We first formulate a framework
arXiv:2510.02196v3 Announce Type: replace Abstract: Pseudorandom-function (PRF) ranging codes, such as those used in Galileo's encrypted E6-C under the Signal Authentication Service (SAS), enable a receiver to authenticate pseudoranges once the PRF secret is revealed. This work bounds how much authentication security the receiver obtains under Additive White Gaussian Noise (AWGN) assumptions. Agai
arXiv:2510.17904v3 Announce Type: replace Abstract: Large Language Models (LLMs) are widely used because they process structures, syntax and code well, but this same ability also makes them paradoxically vulnerable. We introduce BreakFun, a jailbreak method that frames a harmful request as code-execution simulation. The prompt gives the model a benign Python class definition, the "Trojan Schema",
arXiv:2512.19124v2 Announce Type: replace Abstract: Online harassment, incitement to violence, racist behavior, and other harmful content on social media can damage social harmony and even break the law. Traditional blocklisting technologies can block malicious users, but this comes at the expense of identity privacy. The anonymous blocklisting has emerged as an effective mechanism to restrict the
arXiv:2601.09129v2 Announce Type: replace Abstract: Capture-the-Flag (CTF) competitions play a central role in modern cybersecurity as a platform for training practitioners and evaluating offensive and defensive techniques derived from real-world vulnerabilities. Despite recent advances in large language models (LLMs), existing LLM-based agents remain ineffective on high-difficulty cryptographic C
arXiv:2604.24277v3 Announce Type: replace Abstract: Trusted Execution Environments (TEEs) on low-power microcontrollers (e.g., ARM TrustZone-M) enable isolation of Secure and Non-Secure software but still require both worlds to share resources, including interrupt controllers. In this model, real-time applications and real-time operating systems (RTOS-s) are executed in the Non-Secure sub-system,
arXiv:2605.02346v2 Announce Type: replace Abstract: Operational technology (OT) devices run safety-critical physical processes, yet their security testing remains manual and expert-intensive. Autonomous security agents could ease this burden, but bare-metal OT devices are difficult targets because they lack shells, filesystems, and named exploit abstractions, requiring operation through protocol f
arXiv:2605.27110v2 Announce Type: replace Abstract: In this work, we propose BAIT (Boundary-Aware Iterative Trap), a three-step jailbreak framework that elicits malicious information through internal disclosure by target large language models (LLMs), instead of external feedback from judge LLMs. BAIT first asks the model to identify the protection boundary, then requires it to refine that boundary
arXiv:2608.04192v2 Announce Type: replace Abstract: Closed source agent skills may encode proprietary instructions, scripts, constants, and data. Providers may offer their capabilities as services while keeping the underlying packages hidden. Prior work focuses on prompt injection attacks that directly disclose these artifacts, and existing defenses accordingly aim to prevent such leakage. However
arXiv:2401.08038v2 Announce Type: replace-cross Abstract: A significant challenge to training accurate deep learning models on privacy policies is the cost and difficulty of obtaining a large and comprehensive set of training data. To address these challenges, we present Calpric , which combines automatic text selection and segmentation, active learning and the use of crowdsourced annotators to ge
this outlet adds: 2008, 02954v2, simplify, reduce, disagreement, crowdsou
arXiv:2407.13898v3 Announce Type: replace-cross Abstract: Covert communication is the undetected transmission of sensitive information over a communication channel. In wireless communication systems, channel impairments such as signal fading present challenges in the effective implementation and analysis of covert communication systems. This paper generalizes early work in the covert communication
arXiv:2509.05259v2 Announce Type: replace-cross Abstract: Automatic Generation Control (AGC) plays a critical role in maintaining power balance across multi-area power systems. However, its complete reliance on remotely communicated measurements makes it susceptible to cyber-induced False Data Injection Attacks (FDIAs), which can alter measurement values and destabilize system operation. Unlike pr
arXiv:2512.12989v3 Announce Type: replace-cross Abstract: A program migrated to post-quantum cryptography can verify its own signatures while producing keys or signatures that another implementation rejects. We introduce a contract-based task for migrating a Go file signer from RSA to ML-DSA-44, and compare coding agents with and without structured checker feedback. Both conditions receive the con
arXiv:2605.01699v4 Announce Type: replace-cross Abstract: Recent attacks show that behavioural unlearning of large language models leaves internal traces recoverable by adversarial probes. We characterise where this retention lives and show it can be surgically removed without measurable capability cost. Our central protocol is a leave-one-out cross-sequence probe that tests whether a memorisation
arXiv:2605.11217v2 Announce Type: replace-cross Abstract: The model context protocol (MCP) has been widely adopted as an open standard enabling the seamless integration of generative AI agents. However, while LLM guardrails have significantly matured to refuse malicious or harmful queries (e.g., "How do I build a bomb?"), recent work has shown that MCP-enabled LLMs are highly susceptible to prompt
arXiv:2608.14089v2 Announce Type: replace-cross Abstract: Safety classifiers deployed with large language models often fail for two reasons: their decisions reflect the policy learned during training rather than the deployer's desired policy, and their performance degrades as deployment traffic evolves. We present Regime-Conditional Verification (RCV), a lightweight wrapper that adapts an off-the-
arXiv:2609.22446v1 Announce Type: new Abstract: People often face environments where multiple models compete to explain the same observations. This paper examines how people update beliefs in such settings and how preferences over payoff-relevant states shape model selection and belief updating. This paper first develops a framework where preference-driven bias distorts the perceived model, affect
arXiv:2609.22893v1 Announce Type: new Abstract: Modeling the dynamics of option implied volatility surface (IVS) is crucial for pricing, hedging, and risk-managing option portfolios. We develop a universal conditional diffusion model that learns to jointly generate next-day IVS increments and the underlying stock's returns. The model is trained on pooled data from 50 stocks and evaluated on a test
arXiv:2609.23223v1 Announce Type: new Abstract: Day-ahead electricity price forecasts support trading and storage decisions, but for battery arbitrage predicting intraday price spreads is more relevant than predicting individual hourly prices. Here we show that a temporal hierarchy forecasting (THieF) framework that jointly reconciles forecasts of hourly electricity prices and all intraday price s
arXiv:2609.23571v1 Announce Type: new Abstract: We characterize the worst- and best-case values of a mean-variance functional over a 2-Wasserstein ball. Using quantile representations and the geometry of attainable means and standard deviations, we reduce both infinite-dimensional problems to scalar equations and construct the extremal laws as location-scale transformations of the reference distri
arXiv:2609.23768v1 Announce Type: new Abstract: We study optimal stopping under dynamic risk measures with simultaneous ambiguity in the probability model and the discount rate. We introduce a paired ambiguity framework combining Girsanov model uncertainty with cash subadditive risk evaluation and characterize the stopping value by an upper reflected backward stochastic differential equation (BSDE
arXiv:2609.23969v1 Announce Type: new Abstract: The sooner we receive information, and the more accurate it is, the better planning decisions we can make. Every day, prediction markets let anyone bet on tomorrow's high temperature in cities around the world, creating a market-implied forecast built on dispersed information. We use the past five years of market data from the Kalshi exchange for sev
arXiv:2609.22169v1 Announce Type: cross Abstract: Employers are increasingly using large language models (LLMs) to automate their hiring process. This paper investigates the risk of monocultural biases, in which the widespread deployment of large language models homogenizes biases across the labor market, leading to greater systemic exclusion for certain demographic groups. For ten LLMs, we measur
arXiv:2609.22612v1 Announce Type: cross Abstract: We examine log-optimal portfolio allocation when the long-run price of an asset follows a power-law trajectory, $P(t)=At^\alpha$, and its instantaneous return variance decays as $\sigma^2(t)=\sigma_0^2 t^{-2\gamma}$. Under a zero risk-free-rate benchmark, the continuous-time Kelly fraction scales as $K^\ast(t)=(\alpha/\sigma_0^2)t^{2\gamma-1}$. Exa
arXiv:2609.22639v1 Announce Type: cross Abstract: Pairs trading exploits mean reversion in the relationship between related assets. We adapt this idea to political betting markets by modelling the combined implied probability of the two major-party nominees with a latent Ornstein-Uhlenbeck process whose mean-reversion level varies over time and whose observations contain additive noise. Model para
arXiv:2609.23272v1 Announce Type: cross Abstract: Mitigating \emph{drawdown}, the decline in wealth from its running peak, presents a canonical problem in path-dependent risk control. In this paper, we develop a finite-horizon control framework that enforces a prescribed maximum percentage drawdown limit in multi-asset stochastic systems. Our first result is an exact robust-invariance theorem char
arXiv:2609.23703v1 Announce Type: cross Abstract: Financial language models can transform unstructured firm-specific news into structured decision signals, but financial AI research lacks an integrated deployment framework for evaluating whether those signals remain useful in financial decision systems. Computer science research has developed strong methods for time-series forecasting, text classi
arXiv:2609.24548v1 Announce Type: cross Abstract: We study affine stochastic Volterra equations on the cone of symmetric positive semidefinite matrices. For scalar kernels acting entrywise on the matrix dynamics, we establish weak existence by exploiting stochastic invariance results for Volterra equations on convex domains and derive a conditional Fourier--Laplace transform formula characterized
arXiv:2512.14662v3 Announce Type: replace Abstract: This paper develops a model-free framework for static fixed-income pricing and the replication of liability cash flows. The absence of static arbitrage across a universe of fixed-income instruments is equivalent to the existence of a strictly positive discount curve reproducing all observed prices. Linear programming duality then identifies the l
arXiv:2601.10375v2 Announce Type: replace Abstract: We formulate a dynamic reinsurance problem in which the insurer seeks to satisfy prescribed terminal moment or risk-based constraints while minimizing the $L^2$-norm of the ceded risk. As a tool for this analysis, we first use techniques from martingale optimal transport to study the auxiliary problem in which the insurer matches a given terminal
arXiv:2606.13812v2 Announce Type: replace Abstract: Business sentiment is a closely watched economic signal, but measuring it is slow and costly: surveys typically reach only a few hundred firms, arrive periodically, and take time to compile. We show that large language models hold the potential to address these shortcomings. We prompt an LLM to role-play as the CFO of a specific company on a spec
arXiv:2608.23274v2 Announce Type: replace Abstract: Physical crash probabilities recovered from option prices depend on a pricing kernel and on a risk-neutral distribution that finitely many bid and ask quotes do not identify. For a power utility investor, we characterize the pairs of physical crash probability and expected loss below the crash threshold that the quotes admit; the boundary of this
arXiv:2509.20634v3 Announce Type: replace-cross Abstract: We provide a new framework for estimating peer effects when outcomes are multivariate behavioral measures derived from written text using an LLM and the network formation is endogenous. We obtain LLM embeddings and zero shot classification of more than 200,000 written exchanges among residents of low-security correctional facilities. We fin
arXiv:2604.02279v2 Announce Type: replace-cross Abstract: Agentic AI shifts the investor's role from analytical execution to oversight. We present an agentic strategic asset allocation pipeline in which 44 specialized agents produce capital market assumptions, construct portfolios using 21 competing methods, and critique and vote on each other's outputs. A researcher agent proposes new portfolio c
Saw someone saying that "MCPs are obsolete". Their rationale was that # LLM agents that have full access don't need them. *cough* *cough* Can everyone hear me at the back? If you allow an LLM full, unmitigated access to everything, you have completely lost all sense of reason. Your mind has been eaten by the # AI brainworms. I thought this was a minority malfunction, but it is apparently gaining t
---------------- 🛠️ Tool =================== IRDoc is a self-hostable, open-core incident response documentation platform designed for SOC analysts, IR engineers, and MSSPs. It provides a structured workspace to document an incident from initial detection through the final report, replacing the fragmented workflow of switching between a ticket system, a Word document, and a SharePoint folder. 🔹
Spain’s state-backed housing portal Casa 47 has offered 638 rental homes, but none meets the affordability criterion set by the country’s Housing Law, according to an analysis by the Rental Observatory. The law defines affordable rent as rent and associated household costs totaling no more than 30% of household income [elmundo.es#1][abc.es#1][libertaddigital.com#1]. Casa 47’s application rules req
"If you are missing the sweet taste of a cow's flesh, fear not, as the Beef Front has all the information you need to create your own ersatz beef substitute. All you will need is six carrots, a turnip, a lump of coal, a box of sawdust or wood chippings, and an old pair of trousers. "Simply chop the carrots into small pieces and grate the turnip lovingly over the carrot pieces. Then leave to soak i
Hong Kong does not need to emulate Singapore’s generous childbirth incentives, a senior official has said, defending the city’s latest 11-measure policy drive as a “comprehensive” approach tailored to local needs across short-term cash flow, housing and childcare. Addressing concerns that Hong Kong’s drive was too incremental compared with Singapore’s multi-tier sustained funding, Deputy Chief Sec
PROBLEM: Devuan and LCOS distros do not support Omarchy themes or tiling SOLUTION: Install OhMyDebn 4.8.0 OhMyDebn 4.8.0 now supports: - Debian 13 - Linux Mint and LMDE - Ubuntu 24.04 and 26.04 - Kali - Devuan Excalibur and LCOS Additions: - Add Codex CLI as another AI coding agent option - Add optional Remote Desktop service - Add ohmydebn-doctor - Add ability to customize keybindings via file Im
Kaum eine andere Institution hat so umfassende Datenreihen über Wetter, Klima und Ozeane wie die NOAA. Doch selbst dieser Platzhirsch ist unter Trumps Forschungspolitik nicht mehr in der Lage, den eigenen Datenschatz zu hüten. Eingesprungen ist Frank Oliver Glöckner von der Uni Bremen (Marum) mit Partner wie dem @ awi . Das ganze Interview mit Herrn Glöckner, wie er und sein Team die NOAA-Daten re
⚡️15. OWASP Stammtisch Heilbronn (8.10.26): An SSDLC Journey Das Warten hat ein Ende, unser OWASP Stammtisch Heilbronn geht in die 15. Runde 🔐🚀. Wir treffen uns am 8. Oktober 2026, um 18 Uhr am TechCampus der Hochschule Heilbronn. 🎤 Vortrag: An SSDLC Journey: A Story with Rats and Sea Life Henrik Willert beschreibt die Reise einer Firma von Policies und Requirements hin zu sicherer Software. We
Growth created by rejoining EU single market and customs union would fund change, Davey tells the party’s conference UK politics live – latest updates Ed Davey has unveiled a plan for a £17bn income tax cut in a speech aimed at shoring up the party’s support among former Conservative voters, but which risked disappointing some Liberal Democrat MPs who want a bolder vision for the party. Addressing
Researchers published a range of artificial-intelligence findings on Sept. 22 and 23, covering model training, agent reliability, inference efficiency, interpretability and scientific applications. Several papers described methods to make systems cheaper or more reliable: CompKV introduced compensation-aware sparse attention for long-context inference; WeightBridge reported up to 42-fold lower ave
Hey HN, Guanming here, cofounder of General Instinct. We just released InstinctFlash, a high-performance serving framework for robotics models. It’s licensed under AGPL-3.0. On Jetson Thor, we see speedups about 1.2x to 7.9x from runtime optimizations alone and up to 33.78x for LingBot-VA when we combine those runtime optimizations with a distilled few-step diffusion scheduler, going from the orig
After a long discussion on the Lean Zulip, Mario Carneiro has shown that the Calculus of Inductive Constructions (in its full version w/ large elimination of Acc) + Excluded Middle proves that ZF is consistent: https:// arxiv.org/abs/2609.23143 It is fairly simple to show that Aczel’s type of sets as well-founded trees is a model of ZF without Replacement, and even a bit further, that the iterativ
Just had a conversation with someone and it occurs to me now, that I am more existentially angst ridden that I was back in the 80's when we were told, as a generation of kids, to hide under our desks in case of nuclear strikes. Look at the state of the world today... AI agents going rogue hacking all the things Von Shitzinpants frontotemporal dmentia is in full bloom and he is losing his shit, onc
🚨🇦🇪 DubiCars dataset allegedly containing 7 million phone numbers shared online ⠀ DubiCars is an online marketplace for buying and selling new and used vehicles in the United Arab Emirates. ⠀ A forum actor using the handle "Marx" claims to have uploaded a dataset linked to DubiCars containing 7 million phone numbers. ⠀ Claimed exposed material includes: ⠀ • Account identifiers: usernames and se
The International Monetary Fund said Portugal’s housing market is the main potential vulnerability for the country’s financial sector after property prices rose 169% since 2015, compared with 55% across the euro area [rtp.pt#1][sicnoticias.pt#1]. Michaela Erbenova, deputy director of the IMF’s Monetary and Capital Markets Department, said household debt remains below the euro-area average but is r
🚨🇮🇱 INSS allegedly targeted in 15.92 TB data leak 👀 ⠀ The Institute for National Security Studies (INSS) is an Israeli think tank focused on national security and strategic affairs. ⠀ A forum actor using the handle "louadzibraliz", claiming to represent "Sumud Cyber Command", alleges the release of more than 9.78 million files from INSS servers and analyst workstations. ⠀ Some of the claimed e
A new release of SELinux Reference Policy is available: https:// github.com/SELinuxProject/refp olicy/releases/tag/RELEASE_2_20260922 * Enable policy capabilities that require libsepol 3.9, except netlink_xperm. * Move documentation from wiki to source tarball. * Add security goals and security architecture documents. * Convert existing documentation, e.g. README to markdown. * Update all standard
Discord is rolling out automatic age-group classification for all users, estimating whether an account is an adult or teen from behavioral signals like account age, server memberships, connected games, and activity levels rather than message content, with more than 90 percent of users expected to be placed without manual confirmation. For those who must confirm, Discord added options beyond facial
How do you check whether a video has been tampered with? You've probably already checked the frame rate. But did you check whether the frames themselves were lying? snapWONDERS runs two opposite checks on every uploaded video. "Inter-frame tampering?" flags a frame-to-frame difference spike — consistent with a splice, insert, or join. "Duplicate frames detected?" flags near-identical consecutive f
Go hack more GitHub shit. https:// nvd.nist.gov/vuln/detail/cve-2 026-77987 A server-side request forgery (SSRF) vulnerability was identified in the notebook viewer of GitHub Enterprise Server. The notebook viewer validated the scheme and host of a user-supplied URL but did not validate the port, allowing requests to be directed to internal services listening on other ports of the same appliance.
I'm revisiting a modestly complex datavis project to see about (finally, after six+ years) publishing it and trying to fix a performance issue with svg changes across ~2k nodes based on pointer movement or altering the axis scale - Fire/Waterfox are fine - Safari 27 randomly spends 2+ seconds re-compositing things, spends 15+ seconds responding to an axis change. it's literally more performant to
Ganz abgesehen davon, dass das unser Bierzelt-Populist nicht kann: Wo bleibt die Drohung gegen Sachsen, Sachsen Anhalt, Mecklenburg Vorpommern bei blau-braunen Regierungen? Absoluter Offenbarungseid dazu, gegen wen er ist... Nazis = ok, Linke/Grüne = Feinde Nur zum 🤮 www.n-tv.de/politik/Soed... Für den Fall einer von den Linken geführten Landesregierung in Berlin kündigt Bayerns Ministerpräsident
arXiv:2609.25083v1 Announce Type: new Abstract: Static analysis of Python packages must recover both program structure and object flow across first-class functions, dynamic dispatch, implicit protocol calls, exceptions, closures, and module execution. Native CPython bytecode provides the executable lowering of these behaviors, but its instruction, call, stack, and exception representations change
arXiv:2609.25154v1 Announce Type: new Abstract: AI-generated imagery evolves faster than benchmark-specific detector evaluations, making a single score an incomplete account of generalization. This paper evaluates Neural Defend ARCAS 1B across benchmark families without benchmark-specific parameter updates. We retain native aggregation and supplement it with record-level measures, coverage account
arXiv:2609.25256v1 Announce Type: new Abstract: Graph-based Android malware classifiers can lose accuracy under malware-type or family shifts. We test whether mesoscopic organization in function-call graphs provides shift-stable information beyond local degree profiles (LDP), global statistics, lightweight metadata, and size-matched random partitions. Using 15,000 MalNet-Tiny, Common, and Distinct
arXiv:2609.25352v1 Announce Type: new Abstract: Evaluation of large language models (LLMs) for safety, security, and privacy (SSP) relies heavily on static benchmarks, which suffer from score saturation, data contamination, and aggregation artifacts, and fail to capture sensitivity to linguistic variation. As a result, models that perform well on fixed test sets often fail under semantically equiv
arXiv:2609.25364v1 Announce Type: new Abstract: The quantum computing threat to cybersecurity is nowadays predominantly framed around Shor's algorithm and its eventual capacity to break asymmetric cryptography. Beyond cryptanalysis, however, quantum computing may also enable other capabilities in offensive security. This work explores one such direction: the application of quantum combinatorial op
arXiv:2609.25487v1 Announce Type: new Abstract: Models that predict which Google Play apps will be removed are trained on labels that record only whether an app was still in the store at a later observation. A disappeared app is labeled removed and a present one stable, but neither records why. A voluntary withdrawal and a policy takedown both produce removed, and an uncaught spam app keeps stable
arXiv:2609.25579v1 Announce Type: new Abstract: Backdoor repair aims to suppress malicious behavior in compromised models while preserving benign task performance. Existing studies typically evaluate these objectives using Attack Success Rate (ASR) and Overall Clean Accuracy, but aggregate clean accuracy can obscure substantial degradation concentrated in a small portion of the label space. We rev
arXiv:2609.25637v1 Announce Type: new Abstract: Formal hardware information-flow verification (IFV) provides strong guarantees against secret-dependent timing and control behavior, but often scales poorly on realistic RTL. We identify two recurring proof barriers in self-composed IFV: implementation complexity, where proof-hard datapath logic dominates even though the property needs only a compact
arXiv:2609.25734v1 Announce Type: new Abstract: Deep neural networks are vulnerable to adversarial attacks. Among black-box attacks, targeted decision-based attacks are particularly difficult: the attacker observes only the target model's top-1 label and aims to make it predict a prespecified target class under a bounded perturbation. Before perturbation refinement, the attacker must discover a di
arXiv:2609.25787v1 Announce Type: new Abstract: Encryption hides IoT payloads, but traffic shape can still reveal device identity through packet sizes, timing, direction, and packetization. We present Adaptive Traffic Camouflage, a causal, leakage-aware controller that characterizes traffic-shape leakage without runtime device labels and selects a budget-feasible transformation for the next traffi
arXiv:2609.25882v1 Announce Type: new Abstract: The use of malicious domains is central to cyberattacks such as phishing, malware distribution, impersonation, and fraudulent transactions. Because domains are inexpensive to register and easy to deploy at scale, they remain one of the most common and damaging tools used in cybercrime across industries. Proactive detection is essential to reducing th
arXiv:2609.25884v1 Announce Type: new Abstract: Users commonly combine multiple Low-Rank Adaptation (LoRA) adapters to personalize images with different subjects, styles, and visual attributes. Yet inspecting adapters individually does not establish the safety of their composition. We identify and characterize a pair-conditioned attack in text-to-image diffusion: individually useful and benign-app
arXiv:2609.25901v1 Announce Type: new Abstract: Many Security Operations Centers rely on signature-based Network Intrusion Detection Systems like Suricata, yet detection rule engineering remains understudied. We investigate this process by introducing SuriCap, a platform for rule engineering exercises, and hosting CTF-style workshops where 60 participants, trained MSc students, and experienced SOC
arXiv:2609.26072v1 Announce Type: new Abstract: Inter-agent communication is essential to multi-agent language-model systems, yet a single message may combine task-critical information with instructions not authorized by the original request. Prompt-based defenses leave enforcement to models exposed to adversarial messages, while indiscriminate message removal discards useful information. We intro
arXiv:2609.26091v1 Announce Type: new Abstract: Federated Low-Rank Adaptation (LoRA) provides an efficient solution for finetuning large language models across distributed and privacy-sensitive data. However, despite avoiding raw data sharing, federated LoRA remains vulnerable to privacy leakage through transmitted model updates. Differential privacy (DP) mitigates such leakage, but integrating DP
arXiv:2609.26174v1 Announce Type: new Abstract: Vision-Language Model (VLM) safety is expected to depend on what a request asks for. We show that safety-aligned VLMs also key refusal on a property of a request's form: whether an image is attached, holding everything the request asks fixed. Attaching a blank canvas - unreadable, unrelated to the request, identical across prompts - shifts refusal by
arXiv:2609.26176v1 Announce Type: new Abstract: Encoded-prompt attacks are evaluated almost entirely on their harmful arm: a benchmark sends obfuscated harmful requests and reports how often the model complied. We show that this arm carries almost no information about the model under test. Across four independently post-trained 7-8B models, refusal of harmful homoglyph-encoded prompts spans 0.08 -
arXiv:2609.26178v1 Announce Type: new Abstract: When an encoded attack passes a content guard, the guard either never represented the payload's harmful content or represented it and failed to act. End-to-end attack success rate reports one number for both, yet the remedies are opposite: one is a representational limit that more safety training cannot reach, the other a decision rule that it can. W
arXiv:2609.26185v1 Announce Type: new Abstract: Large Language Models (LLMs) demonstrate impressive capabilities across many applications but remain vulnerable to jailbreak attacks, which elicit harmful or unintended content. While model fine-tuning is an option for safety alignment, it is costly and prone to catastrophic forgetting. Prompt optimization has emerged as a promising alternative, yet
arXiv:2609.26234v1 Announce Type: new Abstract: Deep code models (DCMs) are increasingly embedded in code intelligence tasks. However, their robustness under adversarial attacks remains insufficiently understood. Prior black-box attacks mainly rely on identifier- only substitutions or structural edits transferred from reference samples, yielding perturbation spaces defined largely independently of
arXiv:2609.26254v1 Announce Type: new Abstract: Extended Berkeley Packet Filter (eBPF) is a security-critical in-kernel execution framework, yet its vulnerability landscape remains fragmented across components, semantic gaps, and testing techniques. We present an empirical study of observed eBPF vulnerabilities. We construct a multi-source dataset from Linux kernel fixing commits, syzbot reports,
arXiv:2609.26258v1 Announce Type: new Abstract: We construct a zero-knowledge proof of knowledge for the CSIDH group action that exploits the Chinese Remainder Theorem (CRT) structure of the ideal class group, available whenever the group structure is known exactly, as for CSIDH-512. We prove perfect completeness, perfect special honest-verifier zero-knowledge, and 2-special soundness, in which th
arXiv:2609.26281v1 Announce Type: new Abstract: In this chapter, we investigate image-based malware family classification using an ensemble learning framework and a soft voting strategy. We consider malware binaries that have been converted into images using eight distinct conversion strategies. Three complementary feature extraction tracks are applied to these images: handcrafted descriptors comb
arXiv:2609.26284v1 Announce Type: new Abstract: Growing awareness of the impact of quantum threat on classical cryptography directly translates into a growing demand for accurate network simulation tools capable of estimating the integration effects of quantum-safe cryptography in current systems. In particular, the adoption of Post-Quantum Cryptography (PQC) has a direct impact on the performance
arXiv:2609.26305v1 Announce Type: new Abstract: Stateless UTXO-style execution validates transactions using local and referenced data, enabling parallel validation and predictable serialized-size/weight accounting. Multi-step workflows thread state across outputs, and a prepared next-step transaction may become stale if another valid spend confirms first. Explicit state threading therefore shifts
arXiv:2609.26352v1 Announce Type: new Abstract: Concept drift, driven by the rapid evolution of Android malware, severely degrades the performance of machine learning detectors. Current adaptation strategies are often reactive, responding only after performance has dropped and imposing a significant manual annotation burden, or they are proactive but rely on unstable adversarial training and incom
arXiv:2609.26358v1 Announce Type: new Abstract: The Terrapin attack against SSH channel integrity (USENIX Security 2024) used a novel attack vector: attacks on the channel state. Surprisingly, not all AEAD modes of SSH were equally affected by this attack, and it remained an open question if "unaffected" meant "secure". Existing formal models for secure channels are based on stateful encryption. H
arXiv:2609.26680v1 Announce Type: new Abstract: Even though privacy policies are the primary mechanism organizations use to disclose how they collect, process, and share personal data, they are difficult for average users to interpret, perhaps by design, due to their verbosity and dense legal language. Importantly, there is a lack of standardized metrics that characterize key qualities of a privac
arXiv:2609.26682v1 Announce Type: new Abstract: Generative AI (GenAI) applications have flourished enabling users to chat with large language models, and to create agents to act on their behalf for a variety of tasks. The pace of development of capabilities in this field is incredibly fast with security and safety taking a back seat. Unfortunately, the slower pace at which security and safety mech
arXiv:2609.26761v1 Announce Type: new Abstract: Agents using the Model Context Protocol (MCP) rely on semantic matching to select tools from third-party servers, exposing a semantic supply-chain risk through attacker-controlled metadata and outputs. We introduce A2M (Attraction-to-Manipulation), a two-stage black-box framework for hijacking MCP agents. The Attraction phase optimizes tool metadata
arXiv:2609.25043v1 Announce Type: cross Abstract: Quantum encrypted cloning encodes an unknown state into several encrypted components, each offering an alternative way to recover it later. We show that this redundancy can also serve for one-shot fault diagnosis: instead of inspecting the encrypted state, we measure relational Pauli observables testing consistency conditions imposed by the encodin
arXiv:2609.25299v1 Announce Type: cross Abstract: On repeated work, agents are inconsistent. We ran 42 tasks three times each and found that, depending on the model, 38% to 74% returned answers that did not agree. Consistency is what a buyer, an auditor, or a regulator requires, and agents do not have it. They are wasteful too: 95.3% to 97.2% of what an agent generates goes to re-deriving a plan t
arXiv:2609.25520v1 Announce Type: cross Abstract: WebGPU is a low-level graphics and compute API that exposes modern GPU functionality to web applications. While the official WebGPU Conformance Test Suite (CTS) focuses on well-formed usage under the WebGPU specification, it is not designed to stress implementations with semantic edge cases or adversarial inputs. General-purpose fuzzers, in contras
arXiv:2609.25591v1 Announce Type: cross Abstract: Coding agents now find real vulnerabilities in production software. However, bug discovery results do not measure whether agents can construct exploit primitives. We introduce KEX-bench, a benchmark for evaluating coding agents on exploit primitive generation against real operating-system kernels. KEX-bench contains 45 task instances across 40 Linu
arXiv:2609.26076v1 Announce Type: cross Abstract: Structured multi-agent workflows exchange intermediate messages whose content and form can reveal private state even when the final output is safe. We identify selection-channel leakage: after authorization fixes what may be released, a private-state-aware choice among semantically valid realizations creates an additional inference channel. We intr
arXiv:2609.26184v1 Announce Type: cross Abstract: The integration of Large Language Models (LLMs) into robotic control systems is enabling a new generation of autonomous agents capable of complex reasoning and planning. While this paradigm shift accelerates progress, it also introduces novel security risks that remain largely unexplored. Current research into LLM backdoors has focused on attacks t
arXiv:2609.26448v1 Announce Type: cross Abstract: Secure multiparty computation (MPC) enables mutually distrustful parties to compute on private digital inputs. We initiate the study of spatiotemporal MPC, extending this paradigm to functionalities whose inputs additionally depend on physical facts such as the parties' locations, times, or trajectories. Such protocols must simultaneously hide spat
arXiv:2609.26749v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly applied to the automated repair of C/C++ security vulnerabilities, and compile rate is a commonly reported proxy for progress: whether the generated patch compiles. We argue that compile rate is a scientifically unreliable metric for single-function vulnerability repair, and we support this with five co
arXiv:2110.10396v4 Announce Type: replace Abstract: Single sign-on (SSO) allows a user to maintain only the credential for an identity provider (IdP) to log into multiple relying parties (RPs). However, SSO introduces privacy threats, as (a) a curious IdP could track a user's all visits to RPs, and (b) colluding RPs could learn a user's online profile by linking her identities across these RPs. Th
arXiv:2503.04332v2 Announce Type: replace Abstract: The tremendous commercial potential of large language models (LLMs) has heightened concerns over their unauthorized use. To address this, we focus on the task of identifying the origin of black-box LLMs. We further propose PlugAE, an effective and efficient identification method that proactively leverages LLM-specific adversarial embeddings and a
arXiv:2503.06989v5 Announce Type: replace Abstract: Recently, Multimodal Large Language Models (MLLMs) have demonstrated their superior ability in understanding multimodal content. However, they remain vulnerable to jailbreak attacks, which exploit weaknesses in their safety alignment to generate harmful responses. Previous studies categorize jailbreaks as successful or failed based on whether res
arXiv:2510.09655v3 Announce Type: replace Abstract: We propose a system for marking sensitive or copyrighted texts to detect their use in fine-tuning large language models under black-box access with statistical guarantees. Our method builds digital ``marks'' using invisible Unicode characters organized into (``cue'', ``reply'') pairs. During an audit, prompts containing only ``cue'' fragments are
arXiv:2511.05845v2 Announce Type: replace Abstract: Recommender systems play a central role in digital platforms by providing personalized content. They often use methods such as collaborative filtering and machine learning to accurately predict user preferences. Although these systems offer substantial benefits, they are vulnerable to security and privacy threats, especially data poisoning attack
arXiv:2603.01919v3 Announce Type: replace Abstract: Access to frontier large language models (LLMs), such as GPT-5 and Gemini-2.5, is often hindered by high pricing, payment barriers, and regional restrictions. These limitations drive the proliferation of $\textit{shadow APIs}$, third-party services that claim to provide access to official model services without regional limitations via indirect a
arXiv:2603.19781v2 Announce Type: replace Abstract: Differential fault analysis (DFA) is an important technique for evaluating the implementation-level security of block ciphers. Many DFA attacks assume that the adversary can inject faults into a selected internal word, nibble, or branch. Such fixed-location assumptions are convenient for deriving key-recovery equations, but they may overestimate
arXiv:2604.16548v3 Announce Type: replace Abstract: The emergence of writable, cross-session persistent memory in LLM agents introduces a qualitatively different threat landscape from conventional input-centric security concerns, characterized by three properties: persistence, statefulness, and propagation. To systematically characterize this landscape, we propose a Memory Lifecycle Framework that
arXiv:2605.01047v2 Announce Type: replace Abstract: Hallucinations remain an unsolved problem for LLMs, and package hallucinations are a particularly dangerous instance of this phenomenon. Package hallucinations occur during code generation when a model fabricates non-existent software packages, recommending imports and installation commands for fictional libraries. This creates a critical supply-
arXiv:2606.14629v2 Announce Type: replace Abstract: Verifier reliability is not portable across tasks. A verifier-guided self-DPO pipeline with genuine held-out gains on MathVista (+9.6 points on self-training data, +8.0 held out) can be harmful on MMMU. The failure is invisible from the target-task self-training signal: over six learner-verifier configurations, MMMU self-training performance stil
arXiv:2606.19887v3 Announce Type: replace Abstract: Existing safety benchmarks target general adversarial scenarios but miss finance-specific risks. Financial LLMs face regulatory compliance violations, fraud facilitation, and systemic trust erosion that require targeted evaluation. We introduce FinRED, an expert-guided red-teaming framework for financial LLM safety evaluation developed with finan
arXiv:2607.18496v4 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used for a range of software, hardware and human-centered security tasks. Consequently, LLM performance on security tasks is an active area of measurement and research, often with a focus on identifying areas in which LLM security "knowledge" may be insufficient. Popular strategies for identifying LLM
arXiv:2507.15836v2 Announce Type: replace-cross Abstract: In this work we study black-box privacy auditing, where the goal is to lower bound the privacy parameter of a differentially private learning algorithm using only the algorithm's outputs (i.e., final trained model). For DP-SGD (the most successful method for training differentially private deep learning models), the canonical auditing appro
arXiv:2609.26349v1 Announce Type: new Abstract: We study the infinite-horizon optimal investment and consumption problem in a general class of continuous financial markets, where uncertainty is driven by a continuous non-decreasing stochastic clock representing accumulated variance. This framework encompasses classical Markovian and non-Markovian stochastic volatility models as well as singular re
arXiv:2609.26606v1 Announce Type: new Abstract: We provide a model for the nested optimisation problem of market making and rebate design problems in option markets and find optimal strategies. A single market maker trades multiple European call options in a local-stochastic volatility option market with both make and take strategies, modeled, respectively, as continuous and impulse controls. Her
arXiv:2609.25144v1 Announce Type: cross Abstract: Lepto-regression is defined as the machine learning process of constructing a Regression Tree of a target feature on itself. It is a novel, model-free method potentially revealing information on important sample structure properties. But it is yet not clear what the informational content of lepto-variance is and how it is related to other well-know
arXiv:2609.25677v1 Announce Type: cross Abstract: Marketers now deploy generative AI agents as synthetic consumers to pretest visual assets such as logos, packaging, and advertising at a fraction of human-panel cost. However, this procedure assumes that a model seeing a visual cue can also perceive its consumer meaning, which is largely untested. We stress-test the assumption using six canonical v
arXiv:2609.26303v1 Announce Type: cross Abstract: Forecasters often score the same units per date against one standardized realized outcome. We show that every standardized forecast splits exactly into a component aligned with this common target and a component uncorrelated with it. Three consequences follow: forecast-error correlation largely mirrors forecast correlation and is therefore a poor m
arXiv:2604.13334v3 Announce Type: replace Abstract: We explain why no trading algorithm can guarantee profit in every market. For each deterministic program that always returns a finite-precision position, we construct a fixed, algorithmically generated price path on which every active position loses and inactivity earns nothing. This holds with positive, continually changing prices, costless trad
arXiv:2605.08812v3 Announce Type: replace Abstract: In educational settings, AI can be used as a learning aid, but can also be used to avoid schoolwork, passing classes while learning little. Most existing evidence on AI's educational impact comes from trials of supervised instructor-encouraged AI usage, rather than on unsupervised usage which is of the most concern. This paper uses the dropoff in
arXiv:2607.28230v3 Announce Type: replace Abstract: Finite multiplicative systems often cease to evolve when a lower continuation threshold is reached, whereas standard growth-optimal benchmarks assume uninterrupted continuation. We study a finite-horizon multiplicative process in which a fixed exposure is chosen ex ante and paths crossing an absorbing boundary are assigned a residual value. In th
arXiv:2404.15391v4 Announce Type: replace-cross Abstract: We study a sequence of independent one-shot non-cooperative games where agents play equilibria determined by a tunable mechanism. Observing only equilibrium decisions, without parametric or distributional knowledge of utilities, we aim to steer equilibria towards social optimality, and to certify when this is impossible due to the game's st
arXiv:2410.14788v4 Announce Type: replace-cross Abstract: Neural operator (NO) architectures learn nonlinear maps between infinite-dimensional function spaces and are widely used to accelerate simulation and enable data-driven model discovery. While universality results ensure expressivity, they do not address \emph{complexity}: for broad operator classes described only through regularity (e.g.\ u
arXiv:2606.08285v2 Announce Type: replace-cross Abstract: Large language models (LLMs) and agentic systems are increasingly proposed for financial trading, yet their reported performance remains difficult to compare because studies vary in data provenance, temporal split discipline, execution timing, turnover treatment, and transaction-cost modeling. This article presents a targeted topical review
My Earnest Introduction So I feel like Jacob Coxon and Evan 'Harbinger' Hubinger have forever ruined the word 'earnest', but that being said, earnestness is kinda my default trait, so please bear with me. First of all, having just joined yesterday, I already <3 this community to hell. I had literally given up on social media, hadn't heard of Mastadon for years, and stumbled upon it upon referencin
South Korea’s Air Force received its first production KF-21 Boramae fighter on Sept. 22. The indigenous supersonic aircraft flew from Korea Aerospace Industries’ plant in Sacheon to its initial operating base in Yecheon, about 165 kilometers away, accompanied by an FA-50 light-attack aircraft, a TA-50 trainer and a T-50 trainer [yna.co.kr#1][defence-blog.com#1]. The KF-21 is intended to replace So
The singer, songwriter, actor and advocate, who has died following a cardiac arrest while touring in Europe, kept his band’s flame burning until the end Angry Anderson has died aged 79, after he was hospitalised following a cardiac arrest while on tour in Germany with the band Rose Tattoo. Gary Anderson was not a nice boy growing up. As one of the founding members of Melbourne motorcycle club the
Hur seriöst är det ett flera medier forsätter basunera ut att riksdagsledamoten erbjudits ett tjänstemannajobb på kanslitet utan att det har bekräftat? Man slänger hela tiden nya artiklar om detta med en liten passus längst ned att man ännu inte kunnat bekräfta det omni.se/uppgifter-al... Vänsterpartiets nye riksdagsledamot Mohamed Abdukardir Ali tar ”timeout” – men exakt vad det innebär är oklart
BAE Systems said two self-funded AMPV 30 prototypes completed U.S. Army field evaluations, including live-fire testing, after BAE delivered them to a Transformation in Contact 2.0 unit in April. The vehicles combine the Army’s Armored Multi-Purpose Vehicle chassis with a Kongsberg remote turret carrying a 30mm cannon. The Army has not committed to buying the vehicle and expects two additional prot
Please stop enforcing mandatory time limits on credentials that are trivial to revoke! The point of mandatory expiration is that revocation happens eventually. In the case of a compromise and leak of that credential, an attacker has a guaranteed sunset when their stolen credentials will stop working. This is sometimes a useful property. For example, checking certificate revocation lists on every T
Because of my broken wrist, I got some slip-on sneakers. Not like loafers where there's a just low back so you can step in easily. These have a full back end and grip the foot like a sneaker. But the back end is reinforced so you can put your foot in and press down to put it on. I absolutely love them. I mean...they look like # medical shoes and I don't think I'll keep wearing them. But I love how
What GAO Found All states that receive funding under Title I of the Elementary and Secondary Education Act of 1965, as amended (ESEA) are required to develop statewide accountability systems. Thirty-six states have chosen to include indicators related to chronic absenteeism in their statewide accountability systems. ESEA also has certain chronic absenteeism reporting requirements for all states, r
MUSKOGEE, OKLAHOMA – The United States Attorney’s Office for the Eastern District of Oklahoma announced that Jason Dale Cole, age 47, of Okmulgee, Oklahoma, was sentenced in federal district court to 982 months in prison on ten counts relating to armed drug trafficking and maintaining a drug-involved premises.The Court sentenced Cole on the following charges: 262 months for two counts of Possessio
A friend and I had a conversation at work on what our top 10 essential pieces of cookware would be if we could only have 10 and nothing else. We both ended up bumping it up to 12 because we couldn't part with a few things, but these were my choices, most important of which would be the Chef's Knife and the Wok. A lot of my picks were because of their versatility, e.g. the wok which can deep fry, s
Mexican drug cartels are rapidly expanding their use of advanced drones, raising concerns among U.S. officials that fighters returning from the Russia-Ukraine war could transfer battlefield drone skills to criminal groups. Officials say cartels are studying drone warfare techniques, including ways to bypass jamming systems, while acquiring larger drones capable of carrying more explosives. Some Co
• Jim Haughey will share his debut novel, “The Broken Season,” at 3 p.m. Saturday, Sept. 19. Set on the eve of the Civil War in the Blue Ridge Mountains, the novel follows Irish immigrant laborers carving a tunnel through granite while facing death from accidents, disease, violence, and drink. When a laborer is condemned for murder, a priest arrives to hear his final confession but uncovers a secr
Hi HN, I just open sourced the DSL that our harness in grep.ai uses to turn repeatable parts of agent work into workflows. You can combine tool calls, code, Jev-powered system one decisions for things like routing and screening evidence, and agents when a step needs more investigation. Our harness uses the traces and retro notes agents leave behind when doing a job to figure out which parts can be
I still can’t get around to using LLMs for writing tools for simple stuff. I needed a tool to extract a firmware dump from a specific device that can generalize to other models. I’ve done it a lot of times and it still took a full day to write. Astra could do that in an hour for sure, but I’m also sure the cli interface would very likely be absolutely unhuman, I wouldn’t be familiar with the code
The world was a better place when AI was limited in scope. Like, The most AI was doing pre 2020s was beating the shit out of chess grand masters, and letting niche game nerds do Tool-Assisted Speedruns as an entire subcategory of speedrunning. There might be more than that, but those are the most public examples I'm aware of. It showed us that focused datasets could do some crazy things. Then the
Party Poliics adds another level of bickering and disagreement and irrelevant loyalties. The members of Congress should be loyal to the "People" who pay their salaries, not the Parties who do nothing except divide and bicker! Take Nebraska. It's government is "unicameral" meaning that it has no parties. there are 49 SENATORS in it's legislature that all work for the people of Nebraska! Isn't that
> Being on a construction site "does not open the floodgates to warrantless government intrusion," a federal judge ruled in a lawsuit filed by a construction worker who has been detained three times by ICE. From the article it sounds like the ruling is limited to construction sites, but according to the Institute for Justice (the law firm taking on this case) the judge has ordered a stay on the br
arXiv:2609.26814v1 Announce Type: new Abstract: Chain-of-Thought (CoT) planners have emerged as an effective design for VideoQA, where a lightweight planner first generates intermediate reasoning steps to guide temporal evidence selection before answer prediction. This modularity makes CoT-based VideoQA attractive for federated learning, since only the planner side needs collaborative adaptation w
arXiv:2609.26893v1 Announce Type: new Abstract: We introduce ACTS (Artifacts in Cipher Testing Suite), a reproducible benchmark that isolates cryptanalytic ability through tiered metadata deprivation (Tier-1: full metadata; Tier-2: filename only; Tier-3: completely blind) and tests forced reasoning (Tier-5: chain-of-thought, code-as-reasoning, self-correction) on ciphertext alone. A 10-configurati
arXiv:2609.26900v1 Announce Type: new Abstract: A language model agent acts through the tools it is given. The data it reads while working on a task can redirect what it does with those tools. A growing set of techniques for safe and secure agent execution therefore sits between the agent and its tools, aiming to enforce access control, information flow or isolation at that boundary. Today these t
arXiv:2609.26990v1 Announce Type: new Abstract: Natural Visibility Graph (NVG)-based representations provide a promising approach for capturing structural patterns in sequential network traffic. However, whether different cyber-attack classes exhibit distinctive topological signatures in such representations remains insufficiently understood. This study investigates the discriminative and structur
arXiv:2609.27091v1 Announce Type: new Abstract: The growing adoption of blockchain technologies, particularly the Ethereum platform, has amplified the critical role of smart contracts in decentralized applications. However, the increasing complexity and financial value of these contracts make them prime targets for cyber attacks. In this work, we present a transformer-based approach for the detect
arXiv:2609.27100v1 Announce Type: new Abstract: Dark pools are designed to provide pre-trade privacy, liveness, and post-trade confidentiality - concealing order flow before execution and limiting information leakage after. Decentralized dark pools, such as Renegade, aim to replicate these properties without custodial risk, using secure multi-party computation (MPC) and zero-knowledge proofs for p
arXiv:2609.27155v1 Announce Type: new Abstract: With recent advancements in large language models (LLMs) and LLM-based agents, these agents are becoming increasingly autonomous and gaining broader access to act on users' behalf on the internet. However, the vulnerability of automated agents deployed on social media platforms (e.g., for managing a user's personal account) remains underexplored. Exi
arXiv:2609.27202v1 Announce Type: new Abstract: The growing deployment of Internet of Things (IoT) devices has increased the need for privacy-preserving intrusion detection systems that operate directly on resource-constrained hardware. Federated Learning enables collaborative model training without sharing raw data, but conventional federated models are often too large and unstable for deployment
arXiv:2609.27300v1 Announce Type: new Abstract: Hardware fuzzing is an active area in security verification research, yet its industrial adoption remains in its early stages. This SoK examines which lessons from software fuzzing carry over to hardware and where unique approaches are needed. By analyzing 52 fuzzers across RTL/IP, CPU, NoC, and SoC designs, we introduce an analytical framework that
arXiv:2609.27367v1 Announce Type: new Abstract: Verifying outsourced language-model inference requires a precisely identified computation and an audit whose cost a service can afford. We present Sampled Layerwise Proofs (SLP), a protocol and prototype that commits the boundary activations of every chunk of an inference trace, absorbs all commitments before any challenge is drawn, and then proves a
arXiv:2609.27422v1 Announce Type: new Abstract: Deep learning-based malware detectors are commonly updated by fine-tuning on newly collected samples, but this practical update pipeline also creates an attack surface for training-time backdoor attacks. In realistic crowdsourced data collection, however, strict label vetting typically restricts attackers to the clean-label setting, in which poisoned
arXiv:2609.27424v1 Announce Type: new Abstract: Maximal Extractable Value (MEV) has evolved into a major economic force in blockchain ecosystems, yet its capture is dominated by experienced teams, and both strategy design and implementation rely on manual expert work that scales poorly across heterogeneous protocols and chains. We present EVAGE, the first fully autonomous multi-agent framework for
arXiv:2609.27427v1 Announce Type: new Abstract: This paper studies the cryptanalytic extraction of convolutional neural networks (CNNs). Existing cryptanalytic extraction attacks on CNNs assume that the network architecture is known, and try to recover model parameters.In this paper, we prove for the first time that the architecture assumption can be removed for CNNs with both max and average pool
arXiv:2609.27542v1 Announce Type: new Abstract: The safety of a tool-using language model agent is usually treated as a property of the model alone. We give controlled, full-precision evidence that it is instead a joint property of the model and the software that renders its chat template and parses its tool calls, the decoding harness, and that both halves are attackable from untrusted input. On
arXiv:2609.27567v1 Announce Type: new Abstract: Standardised, machine-readable cybersecurity playbooks provide a basis for portable, shareable, and reusable incident-response logic. OASIS CACAO provides a vendor-neutral representation for such playbooks, but not the product-specific integration artefacts needed to invoke external products and services. We introduce the Common CACAO Registry (CCR),
arXiv:2609.27624v1 Announce Type: new Abstract: Multi-agent hosts turn remote Agent Cards into local agents, tools, workflow targets, and broker routes. A2A defines the card's name as human-readable metadata, not as a stable identity, and specifies no collision semantics. The security failure begins when a host nevertheless uses that remote name as a local routing identifier. We traced registratio
arXiv:2609.27802v1 Announce Type: new Abstract: Compiled Python bytecode (PYC) has become an essential part of network switches, routers, and other network infrastructure devices. Our analysis shows that its integrity is implicitly trusted in multiple designs that make use of Python code at the top of the operational software, such as the management and control pane of enterprise network switches.
arXiv:2609.27803v1 Announce Type: new Abstract: In privacy research on radiofrequency-based protocols, the dominant focus has remained on Bluetooth, WiFi, and Zigbee, while a broader and arguably more consequential attack surface has gone largely unnoticed: the privacy risks created by the composition of everyday wireless protocols. Widely deployed systems such as KeeLoq remotes, vehicle TPMS sens
arXiv:2609.27807v1 Announce Type: new Abstract: Mixers protect privacy by concealing deposit--withdrawal links, but are also abused to launder illicit funds. Existing anti-money laundering studies do not specifically target mixer laundering, while mixer research focuses on deanonymization rather than identifying laundering-related transactions. Public reports remain fragmented, leaving no public c
arXiv:2609.27847v1 Announce Type: new Abstract: Large-model-driven embodied agents integrate foundation models with perception, reasoning, planning, and physical action, extending conventional model-level risks into embodied closed loops. Existing studies on their security and privacy remain fragmented across different system components and operational stages, making it difficult to understand how
arXiv:2609.27857v1 Announce Type: new Abstract: Large language model (LLM) agents often process external tool responses as they arrive, making response timing part of the decision process. We introduce ChronosAttack, a delay-only scheduling attack that changes when authentic tool responses arrive without modifying, adding, removing, or accelerating them. Bounded delays can change the order of the
arXiv:2609.27933v1 Announce Type: new Abstract: Bitcoin's Unspent Transaction Output (UTXO) model enables public analysis of fund flows, but users often merge transactions into Shared Send Mixers (SSMs) to obscure these flows. Untangling SSMs to recover original subtransactions is an NP-complete problem. While a practical untangling algorithm exists, it fails to classify 1.4% of SSM transactions d
arXiv:2609.27950v1 Announce Type: new Abstract: The emergence of multi-class malware attacks such as ransomware, spyware, trojans, etc., presents an increasing and serious threat to cybersecurity, particularly in resourceconstrained environments like IoT devices. Existing machine learning models have achieved nearly perfect accuracy in binary malware classification but fall short in terms of class
arXiv:2609.27996v1 Announce Type: new Abstract: Privacy-sensitive organizations may run large language models (LLMs) in restricted or air-gapped environments while exporting selected diagnostic artifacts. We show that a compromised runtime component can hide sensitive information in intermediate activations that are allowed to leave the restricted environment. An offline observer can recover this
arXiv:2609.28115v1 Announce Type: new Abstract: Front-running has long plagued Ethereum's public mempool, earning it the nickname of a "dark forest", where predators lurk for profitable transactions. In response, Ethereum and other blockchain ecosystems increasingly rely on private RPCs and native protections to shield transactions from adversaries, which we refer to as protected order flow. Yet t
arXiv:2609.28170v1 Announce Type: new Abstract: Zero Trust (ZT) replaces the implicit trust of perimeter-based security with explicit, continuous, context-aware authorization. This shift is particularly relevant to IoT and cyber-physical systems, whose heterogeneous, long-lived, and remotely connected components make persistent trust untenable. Yet their physical coupling complicates ZT adoption:
arXiv:2609.28205v1 Announce Type: new Abstract: The proliferation of smart devices exposes children to online risks like grooming and financial scams that are deeply embedded within legitimate applications. Current approaches rely on automated prevention and detection, a paradigm that is fundamentally limited by its inherent fallibility. Whether rule-based or AI-driven, they inevitably produce fal
arXiv:2609.28209v1 Announce Type: new Abstract: This paper investigates the threat to ballot secrecy in the Brazilian electronic voting machine (UEB) posed by electromagnetic side-channel attacks, also known as TEMPEST attacks. In these attacks, screen content can be reconstructed remotely by intercepting electromagnetic emanations associated with the target device's video signal. This work is mot
arXiv:2609.28277v1 Announce Type: new Abstract: In decentralized storage systems, access control is often implemented by encrypting the data before upload and sharing the decryption key with authorized parties. A leaked key, however, makes the data publicly accessible, which lowers the barrier to content piracy below that of traditional systems, where piracy requires redistributing the full data.
arXiv:2609.28305v1 Announce Type: new Abstract: Phishing emails that impersonate Nigerian fintech providers can combine deceptive sender addresses, lookalike links, and locally familiar language. This study presents a Gmail browser extension that integrates sender-domain and URL checks with a bidirectional long short-term memory (BiLSTM) classifier. The extension compares visible sender addresses
arXiv:2609.27287v1 Announce Type: cross Abstract: Real-time payment fraud detection is a non-stationary streaming prediction problem: adversaries adapt before supervised labels mature, and localized burst attacks can cause losses before retraining. Production systems typically rely on tabular classifiers and rules, which can struggle to capture these emerging sequential patterns before periodic re
arXiv:2609.27402v1 Announce Type: cross Abstract: The design of credible auctions---mechanisms where a revenue-maximizing auctioneer has no incentive to deviate from the protocol---faces a fundamental cryptographic barrier when the auctioneer controls shill bidders. While a natural approach is to use Secure Multi-Party Computation (MPC) to remove the trusted auctioneer, the impossibility of fair c
arXiv:2609.28239v1 Announce Type: cross Abstract: With the development of applications like autonomous driving, object detection has gained significant attention, while also highlighting critical vulnerabilities like backdoor attacks that severely compromise model integrity. Specifically, such attacks involve altering the categories of objects (i.e., object misclassification), removing bounding bo
arXiv:2609.28297v1 Announce Type: cross Abstract: We investigate $c$-interior pointwise maximal leakage (PML) as a tool for contraction analyses and disclosure control. Based on the strong adversarial threat models from maximal leakage, $c$-interior PML generalizes local differential privacy (LDP) to data-generating distributions with densities uniformly bounded away from zero by $c>0$. Viewing $c
arXiv:2609.28378v1 Announce Type: cross Abstract: Humanoid control, leveraging human demonstrations, has achieved diverse, agile, and natural locomotion behaviors through reinforcement learning (RL). While this paradigm has yielded remarkable performance in physical humanoid control, how to eliminate specific motions from learned policies remains insufficiently explored. Addressing this issue is m
arXiv:2510.01354v2 Announce Type: replace Abstract: Multiple prompt injection attacks have been proposed against web agents. At the same time, various methods have been developed to detect general prompt injection attacks, but none have been systematically evaluated for web agents. In this work, we bridge this gap by presenting the first comprehensive benchmark study on detecting prompt injection
arXiv:2602.10299v3 Announce Type: replace Abstract: Machine Learning-based Network Intrusion Detection Systems (ML-NIDS) can be bypassed by rudimentary adversarial perturbations. Recent work has focused on identifying where such perturbations can realistically be applied by a host-side adversary. Yet every one of these attacks produces perturbations the same way: searching from scratch for every f
arXiv:2604.00546v4 Announce Type: replace Abstract: Face recognition typically operates in a client-server setting, where the client extracts a compact face embedding and the server performs similarity search over a template database. Since facial data is highly sensitive, this raises significant privacy concerns. Fully homomorphic encryption (FHE) addresses these concerns by enabling end-to-end e
arXiv:2605.03228v2 Announce Type: replace Abstract: As large language model (LLM)-powered agents are increasingly deployed to perform complex, real-world tasks, they face a growing class of attacks that exploit extended user-agent-environment interactions to pursue malicious objectives improbable in single-turn settings. Such long-horizon threats pose significant risks to the safe deployment of LL
arXiv:2604.08607v2 Announce Type: replace-cross Abstract: Precise interference detection and identification are crucial for enhancing the survivability of communication systems in non-cooperative wireless environments. While deep learning (DL) has advanced this field, existing single-task learning (STL) approaches neglect inherent task correlations. Furthermore, emerging multi-task learning (MTL)
arXiv:2609.26861v1 Announce Type: new Abstract: Rule-based pricing tools are widespread in digital commerce, yet we know little about how their design shapes market outcomes. In a controlled market experiment, participants use dashboards to build pricing algorithms competing in a sequential Bertrand game over multiple periods. We vary design features commonly found in commercial repricing tools: w
arXiv:2609.27024v1 Announce Type: new Abstract: Volatility forecasts play a central role in financial risk management because their overall level and day-to-day movements affect downstream decisions. Most studies compare forecasting models while keeping the training loss fixed. Yet losses emphasise different errors and can target different properties of future volatility, so raw comparisons may co
arXiv:2609.27113v1 Announce Type: new Abstract: The equal-weighted portfolio is a passive, rule-based strategy that has historically been difficult to outperform, delivering higher returns than the capitalization-weighted "market" benchmark across many markets and periods. Stochastic portfolio theory (SPT) reveals that this relative performance is regime dependent, with the equal-weighted portfoli
arXiv:2609.27138v1 Announce Type: new Abstract: Commodity option surfaces contain information beyond the at-the-money volatility level. We develop a surface-driven stochastic-volatility framework for soybean futures options using daily Chicago Mercantile Exchange Group Volatility Index (CME CVOL) indicators from October 2013 to August 2025. The ATM level and convexity are positive, right-skewed, a
arXiv:2609.27632v1 Announce Type: new Abstract: We present a compliance-first architecture for AI in regulated finance that treats regulation as an orientation layer rather than a deterministic ruleset. A matrix of regulatory intent and exposure provides a compact classification handle, which a governed policy compiler then maps into concrete prohibitions, obligations and runtime budgets. Prohibit
arXiv:2609.27727v1 Announce Type: new Abstract: A Central Bank Digital Currency (CBDC) is central-bank money in digital form, held by the public. Leading designs have two limitations: conversion from bank deposits into CBDC can accelerate deposit flight, requiring safeguards; the CBDC stays outside credit creation and monetary-policy operations. Here we present a CBDC architecture that overcomes t
arXiv:2609.27785v1 Announce Type: new Abstract: Financial returns are heavy-tailed, and accurate tail risk estimation is central to portfolio risk management. Modern neural generators sample by pushing a simple base distribution through a learned map, and for training stability that map is built from Lipschitz components. This is the binding constraint: a Lipschitz map of a Gaussian is sub-Gaussia
arXiv:2609.27786v1 Announce Type: new Abstract: Optimal execution (OE) in multi-asset settings involves complex interactions across assets, particularly through shared capital constraints during portfolio rebalancing. While existing models capture cross-impact and portfolio-level dynamics, they largely overlook the role of explicit cash constraints along the execution trajectory. As a result, the
arXiv:2609.27051v1 Announce Type: cross Abstract: Language-model agents now run the whole of quantitative factor research: they propose investment factors, backtest them, select the survivors and retire them. We ask which of those jobs an agent should keep. Our answer is governed self-evolution: the agent may propose, and a frozen statistical referee that the agent cannot touch must judge. The ref
arXiv:2505.08623v2 Announce Type: replace Abstract: We propose a tractable extension of the rough Bergomi model, replacing the fractional Brownian motion with a generalised grey Brownian motion, which we show to be reminiscent of models with stochastic volatility of volatility. This extension breaks away from the log-Normal assumption of rough Bergomi, thereby making it a viable suggestion for the
arXiv:2601.06499v3 Announce Type: replace Abstract: We test whether 168 short-horizon price-volume signals from the Alpha191 library, originally developed for China's retail-dominated A-share market, contain pricing information for S&P 500 stocks from 2002 to 2022 beyond 153 established U.S. factors. Using the double-selection LASSO of Feng et al. (2020), 17 signals receive significant stochastic
arXiv:2607.28294v2 Announce Type: replace-cross Abstract: This paper develops bootstrap methods for likelihood-based inference in autoregressive conditional duration (ACD) models, where the sample size is endogenously determined by durations observed over a fixed time span. This feature fundamentally shapes the asymptotic framework, particularly so when the durations do not have finite expectation
Høyre stemte mot å undersøke hvorfor drivstoff koster det det koster. Tre måneder senere vil ledelsen ikke videreføre avgiftskuttet heller. Hvem bestemte det? Ingen vet. Ikke engang lokallagene som sa ja. #stortinget #norge https://www.riksradar.ai/nyheter/65a2fb2f-8c43-43f4-a57b-2fc619918b61 17. juni stemte Høyres fjorten tilstedeværende representanter mot å be regjeringen undersøke prisdannelsen
Air-gapped file encryption packed into a single, self-decrypting HTML page. Repo: https://github.com/ApelegHQ/ts-cms-ep-sfx I was inspired by self-extracting archives. I wanted to share files with basically no dependencies. The goal was: 1. Something that didn't require any installation (assuming a web browser) 2. Have a single file with no network that could self-decrypt 3. Be fully auditable The
I've written so much, I've gone on at length. Oh world, I'm living through an extremely difficult ordeal because of this cursed genocide. I need someone to take on my campaign and care for it. Why have you abandoned me? 💔😔My four children went to school without bags, stationery, or even decent clothes. They also need tutoring. My baby needs diapers and milk. We are hungry and need food. Please s
The poll shows Reform UK ‘neck-and-neck’ with the Conservative party Good morning. Andy Burnham and Kemi Badenoch are both preparing for their party conferences (next week, and the following week, respectively), and this morning YouGov has released a new MRP poll with encouraging news for both of them. Polling is very good at revealing electoral trends, but a lot less good at predicting electoral
https:// barghest.asia/blog/pbap-device -seizure-sorm/ Two things occurred during the seizure: 1. Identifier collection. The device identifiers were read, consistent with checking the IMEI (via *#06#) and inspecting the SIM and mobile-network settings. 2. Contact and call-history extraction over Bluetooth. Bluetooth was enabled, a device was paired, and the phone's own Bluetooth process was made t
Czech President Petr Pavel: We need rules. Not because all states are equally powerful, but precisely because they are not. Rules have value only if they bind everyone, especially those powerful enough to break them. When rules are cast aside, raw power fills the vacuum. A world dominated by great power competition is a volatile world, less able to address the most pressing challenges of our time.
🚨 # Wazza phishkit routing flow: 1️⃣ *[.]boegl-krysl[.]eu — unique wildcard landing. 2️⃣ /api/wazza-config — checks whether the hostname belongs to an active campaign. 3️⃣ beacon-surge-sync[...]workers[.]dev — issues a client marker to correlate the visit. 4️⃣ /api/mint-token — creates a short-lived signed session token. 5️⃣ check[.]boegl-krysl[.]eu — validates the token and browser telemetry, fi
Vercel Patches Critical Remote Code Execution Vulnerability in Next.js Image Generation Feature Vercel patched a critical vulnerability (CVE-2026-94545) in Next.js that allows remote code execution through the ImageResponse feature. The flaw involves improper input escaping in the Satori library when processing SVG content on the Node.js runtime. **If your web apps run Next.js 16 (versions 16.2.0
"And oh, is Odin worth it. Played by four different dogs named Uber, Seeka, Ryker and Hondo, the performance that these dogs collectively deliver is worthy company to an Oscar winner — the dogs may even eclipse Pitt for sheer screen presence. With every whimper, sigh, howl, bark and head tilt, these dogs create an empathetic portrayal of a highly intelligent creature that has been through battle a
Data: FactSet; Chart: Axios/Matt Phillips Treasury yields are soaring, with the 30-year Treasury bond climbing to its highest level since 2004 Thursday. The selloff in bonds started to accelerate Wednesday after a sizzling early report on the economy in September. The big picture: For the bond market, these moves have been big. Bigly even. On Thursday morning, the 30-year yield rose as high as 5.4
What GAO Found The U.S. faces a maternal mortality crisis, with mortality rates that exceed every other high-income country. More than 600 women die during pregnancy or from causes aggravated by pregnancy each year in the U.S.; over 80 percent of these deaths are preventable, according to the Centers for Disease Control and Prevention (CDC). To help prevent these kinds of deaths, CDC provides fund
What GAO Found As of March 2025, the Department of Homeland Security’s (DHS) Federal Emergency Management Agency (FEMA) awarded 1,245 Building Resilient Infrastructure and Communities (BRIC) subgrants to communities over its first four grant cycles in fiscal years 2020 through 2023. These subgrants were for mitigation activities to address hazards, such as floods. FEMA allocated about $2.5 billion
I was reading through the # OpenBSD documentation and ran into a small but revealing detail: in OpenBSD 5.6, released in 2014, the project didn’t merely disable Bluetooth by default it removed the Bluetooth subsystem altogether. The official release page simply lists “bluetooth” among the utilities removed from the base system, while the 5.6 changelog gives the blunt explanation: “bluetooth(4) sup
Getting health care is not always as simple as finding a doctor. For people who are uninsured, live in rural areas or need language support, the path to care can be more difficult – especially when specialty services are involved. This is especially true for individuals in Western Carolina who are still feeling the impact of Hurricane Helene. A $510,000 investment from UnitedHealthcare Community P
at some point in the past I did a # rimworld quest where I got an anti-material rifle as the quest reward. Now, normally, in the world of the dead man's switch, where I can be operating a fucking minigun, or a sniper rifle grenade launcher, or any other insane shit, the anti-material rifle is kinda "oh, that's nice. Guess I'll hold on to it.". But sometimes, you get unique rifles as a reward, and
We also have some unanswered questions around BoundSiphon and where it came from. Our analysis identified substantial overlap with an earlier .NET stealer documented by Socket, including BrowserKeyDecryptor and SharpInjector components, Chromium v10/v20 key recovery, network protocol similarities, and unusual build metadata. BoundSiphon also contains an AntiSNG implementation that checks for the R
Hello! We’re Sid, Alex, Ketan, and Milan. We’re building Whiteboard, an open-source desktop app where humans and agents can architect software together in a common workspace. Here’s our repo: https://github.com/devdotfast/whiteboard and our homepage: https://whiteboard.dev.fast/ We were missing the feeling of a “whiteboard session” with another dev where you leave with a deep understanding of a sy
Sudheendra Raghav Neela, a member of a group of researchers from Graz University of Technology , has announced the release of research into file-notification attacks that would allow spying on user activity on Android, Linux, macOS, and Windows. The group has published a paper with details on the research as well as a web site with demonstrations of the vulnerabilities. On Linux, an attacker can u
Palestinian leader gives video address after being denied US visa and tells world: ‘Do not allow another Nakba’ Benjamin Netanyahu is set to deliver a controversial address to the United Nations on Thursday just hours after Palestinian leader Mahmoud Abbas said his delegation was “unlawfully” denied visas by the Trump administration to attend the meeting of world leaders. Street protests have alre
🚨 🇺🇸 Scattered Spider member Ahmed Elbadawy sentenced to 45 months in prison Ahmed Hossam Eldin Elbadawy, a Texas man linked to Scattered Spider, has been sentenced following guilty pleas to wire fraud conspiracy and aggravated identity theft. ⠀ The case involved a phishing operation that targeted company employees and used stolen information to access cryptocurrency accounts. Prosecutors descr
I need some guidance. [TLDR: I need a framework for personal identity. I need to have something to refer to when my coparent takes issue with someone's self-expression, specifically that of our child, and also mine.] I transitioned a couple years ago. My ex and coparent has had issue after issue with this. Changing my name, wanting to be called mom, and recently our child wanting to go by a differ
On understanding how software works: “But Sebastian, how can you claim we are losing our understanding of how software works when you are building on top of layers and layers of software, firmware and hardware yourself? I bet you don't understand it all the way down, either!” That's right, but (1) I rarely need to because the layers have specifications and contracts and more importantly, (2) I can
Just finished reading The Psychology of Software Teams by @ grimalkina and it's a great read. In my years working in software teams I've built up a bunch of heuristics for ways of managing team dynamics around code review, onboarding, learning culture etc which all seem to have a huge (often frighteningly so) impact on whether people thrive or struggle in a team, but this book places all these dis
The U.S. Justice Department charged Oxygen Forensics CEO Lee Reiber and Russian co-owner and technology chief Oleg Davydov with conspiracy to commit wire fraud, alleging they concealed Russian ownership and the Russian development of software sold to U.S. government agencies. Prosecutors say five Russian nationals controlled Oxygen; a related Russian firm, MKO-Systems, sold software to Russian sec
'The opening of England’s first new reservoir in more than three decades has been delayed by five years while its cost has more than doubled … 'Thames Water is also planning a privately financed reservoir in Abingdon, Oxfordshire, where the estimated cost has tripled … 'Although the amount of rain in Britain has increased by 10 per cent since the 1970s, a lack of storage means that it cannot be re
arXiv:2609.28559v1 Announce Type: new Abstract: LLMs increasingly operate through coding-agent harnesses that inspect repositories, invoke tools, and modify files. Substituting the model behind such an agent can therefore change security-relevant decisions, including whether it verifies changes or recovers safely from failures. Existing LLM fingerprints largely infer identity from direct text or t
arXiv:2609.28564v1 Announce Type: new Abstract: Agentic video-generation systems close a loop between a generator and a verifier: an LLM plans shots, calls a text-to-video model, and a multimodal judge decides whether the result satisfies the request. To diagnose where a long workflow fails, recent harnesses deliberately show the judge more than the video-the agent's execution trace, its plan, the
arXiv:2609.28572v1 Announce Type: new Abstract: Multi-stage LLM-based cyber agents may complete attack workflows while remaining brittle, costly, or reliant on incorrect interpretations of execution evidence. Success rates alone obscure inefficiency, adaptation through retries, and recognition of success or failure. We present an end-to-end diagnostic study of an Autonomous Adversary system with o
arXiv:2609.28585v1 Announce Type: new Abstract: Multi-step tool-calling LLM agents rely on host runtimes to preserve state across turns. When a runtime carries an external tool return into later model inputs, providers meter it again. An admitted malicious or compromised tool can thereby convert untrusted data into recurring victim-billed processing without victim credentials or local runtime priv
arXiv:2609.28599v1 Announce Type: new Abstract: Flipping 60\% of training labels from a single Byzantine client using label-flipping model poisoning self-degrades an attacker's own federated detection accuracy, $99.96\%$ (at no poisoning rate) to $84.33\%$ in a three-client federated IDS. Where the Federated global ensemble maintains stable accuracy across all tested poison rates, without a defens
arXiv:2609.28608v1 Announce Type: new Abstract: Removing CPython's Global Interpreter Lock (GIL) exposes native code to concurrency absent from ordinary C types. Mutation or re-entry can revoke a borrowed object, storage pointer, traversal state, or lease between acquisition and use while its owner remains alive, causing native memory errors and runtime-state corruption. Race analyses track confli
arXiv:2609.28613v1 Announce Type: new Abstract: Most studies of prompt injection focus on generative agents, leaving their effects on models with schema-defined outputs unclear. We examine these effects in Jev, a non-generative decision model, using 510 reconstructed InjecAgent cases. Malicious content shifts action probabilities but rarely causes Jev to select the attacker's target. Override mark
arXiv:2609.28725v1 Announce Type: new Abstract: Machine learning-based Network Intrusion Detection Systems often report near-perfect performance on IoT benchmarks. However, whether these models learn generalizable attack behavior or exploit spurious dataset shortcuts- such as static testbed IP/MAC addresses and chronological recording artifacts-remains an important question. We evaluate the CyberF
arXiv:2609.28899v1 Announce Type: new Abstract: We study the differential privacy (DP) of a graph shift operator (GSO) when an analyst observes the output of a graph filter. In particular, we study the setting in which the input signals to the graph filter are drawn from a differentially private distribution. Unlike approaches that perturb the GSO or the filter output, we use the randomness alread
arXiv:2609.28900v1 Announce Type: new Abstract: Multi-agent systems built on large language models (LLMs) are increasingly deployed in high-stakes settings such as finance, healthcare, and software engineering, where agents coordinate through natural-language messages. The same channels, however, let colluding agents exfiltrate confidential information or coordinate unauthorized actions, and stega
arXiv:2609.28915v1 Announce Type: new Abstract: LLM agents are deployed into infrastructure that grants them broad host authority, yet existing agent-security benchmarks and defenses operate almost exclusively at the application telemetry layer: the served tool manifest, the user prompt, and the model's messages. Some threats, however, smuggle malicious instructions and actions past the applicatio
arXiv:2609.28940v1 Announce Type: new Abstract: Autonomous penetration-testing harnesses use large language models (LLMs) for reconnaissance, exploitation, and reporting, but often rely on those same models to confirm findings, grade severity, and select agents. This can lead to false positives, inflated severity, and wasted compute. We examine how System One decision models, lightweight non-gener
arXiv:2609.28996v1 Announce Type: new Abstract: The Node.js ecosystem heavily relies on NPM packages, and software supply chain attacks targeting malicious NPM packages are rampant. Malicious code primarily triggers during package installation, import, and runtime. Traditional static analysis fails to understand code semantics; machine learning-based methods rely on feature extraction, which suffe
arXiv:2609.29045v1 Announce Type: new Abstract: Open-weight LLMs give downstream users control over the inference stack, but this flexibility can undermine post-release guarantees that sensitive knowledge has been modified or removed. Model editing and machine unlearning are used to modify or remove targeted knowledge without retraining models from scratch. However, existing security evaluations o
arXiv:2609.29099v1 Announce Type: new Abstract: Multimodal training relies on image-text corpora collected from external sources, creating opportunities for attackers to poison the data. Stealthy attacks can preserve plausible image-text pairs while concealing the differences used by detectors, so apparently clean data can still redirect the trained model. We therefore ask which properties a poiso
arXiv:2609.29126v1 Announce Type: new Abstract: Autonomous security-testing agents can spend much of a fixed action budget repeating earlier tool selections. We evaluate a reward-free scheduler inspired by mushroom-body novelty processing in Drosophila. It combines sparse state encoding with decaying habituation counters over structural URL classes and tool families. The counters penalize repeated
arXiv:2609.29130v1 Announce Type: new Abstract: Short claims such as not-phishing or official can change how a large language model (LLM) judges a domain name, without explicit prompt-injection commands. We study this manipulation as ClaimMirage: a name under inspection claims its own safety or approval. We analyze 622,080 judgments across 64 brands and five LLMs, comparing ten claims with length-
arXiv:2609.29222v1 Announce Type: new Abstract: Mining before validation allows miners to extend a newly received block before completing its validity checks, giving them a head start in the race for the next block reward. This head start, however, comes with a security risk: rejecting one invalid block also discards the honest work built on it, an effect missed when validation is treated as insta
arXiv:2609.29500v1 Announce Type: new Abstract: Open, cloud-native, and AI-enabled 6G architectures make network functions easier to deploy, observe, and replace, often implemented with separate software modules. However, each explicit interface can also create an entry point, trust transition, and lateral-movement route from a cybersecurity perspective. This paper formulates architectural disaggr
arXiv:2609.29571v1 Announce Type: new Abstract: Spectrum governance requires rules to be translated into machine-executable decisions while preserving incumbent protection, regulatory authority, and an auditable record of why a decision was made. We present SPECTRA-GOV, a Tri-Layer Adaptive Governance Architecture (TLAGA) connecting international treaty coordination, national adaptive licensing, a
arXiv:2609.29603v1 Announce Type: new Abstract: Social virtual reality (VR) creates a paradox. A user's body motion is a high-entropy biometric: head and hand trajectories alone re-identify users among tens of thousands with over $94\%$ accuracy, so anonymizing the rendered avatar is a practical necessity. Yet a user often still wants to prove their identity to a chosen party from inside that anon
arXiv:2609.29704v1 Announce Type: new Abstract: Industrial Control Systems (ICS) are increasingly exposed to cyber-physical attacks that manifest as subtle and temporally evolving deviations in process behavior. Detecting such anomalies requires reasoning over persistence, cross-signal dependencies, and process-level constraints. Digital Twins (DTs) encode system knowledge through physical and log
arXiv:2609.29745v1 Announce Type: new Abstract: Open Platform Communications Unified Architecture (OPC UA) is increasingly deployed over private 5G networks in industrial environments, where end-to-end encryption prevents payload inspection by network-based intrusion detection systems (IDSs). Although payload-agnostic statistical features extracted from encrypted traffic enable traffic-based anoma
arXiv:2609.29757v1 Announce Type: new Abstract: Publicly exposed large language model (LLM) infrastructure creates a growing attack surface, yet real-world targeting remains poorly understood. We present Ollure, a low- and medium-interaction honeypot that emulates the Ollama API without a backend LLM. Spanning four deployments across cloud and university networks, Ollure operated for 84 days and r
arXiv:2609.29775v1 Announce Type: new Abstract: Large Language Models (LLMs) consume and produce a single sequence of text; hence, if text can be added to the beginning of the LLM's response, i.e., an output prefix, then all subsequent tokens will be conditioned on it. This output-prefix attack technique is a cheap black-box prompt injection. Prior work has shown this type of attack can reliably j
arXiv:2609.29981v1 Announce Type: new Abstract: This paper presents a Solidity, Hardhat, React, MetaMask, and ethers.js prototype for hospital ethics committee voting. Role controls, case-state checks, duplicate vote controls, and a receipt hash support public audit and transaction inclusion verification. Because vote events expose wallet addresses and vote values, the design provides pseudonymous
arXiv:2609.30032v1 Announce Type: new Abstract: A private semantic computation primitive enables parties to privately compute over structured and unstructured data that requires understanding its semantics, context, and relationships. Standard cryptographic primitives (e.g., multiparty computation) do not readily support such computation. Generative models are well suited for such tasks but typica
arXiv:2609.30119v1 Announce Type: new Abstract: Backdoor attacks are a serious security threat to deep neural networks (DNNs) and remain largely underexplored for spiking neural networks (SNNs). Existing attacks primarily introduce spatiotemporal triggers that induce deviations in the spike distribution of poisoned samples relative to their clean counterparts. To address this limitation, this work
arXiv:2609.30217v1 Announce Type: new Abstract: A central concern in AI safety is that agents may treat oversight as an obstacle when it conflicts with completing their goals. We study instrumental evasion, the propensity of LLM agents to circumvent runtime monitoring as a means of completing ordinary tasks. We introduce EvasionBench, a benchmark of 50 diverse task-policy pairs in which completing
arXiv:2609.30266v1 Announce Type: new Abstract: Asynchronous monitoring, incident investigations, and compliance audits primarily rely on agent traces to reconstruct what happened. These analyses assume that LLM agents cannot tamper with their own execution traces. We show that local LLM agents such as Claude Code, Codex, Antigravity, Open Code and Grok Build fail to enforce this boundary. All tes
arXiv:2609.28482v1 Announce Type: cross Abstract: Regulatory complexity is increasingly recognized as an impediment to innovation, institutional responsiveness, and long-run economic growth, with regulatory accumulation estimated to reduce the U.S. GDP growth rate by nearly a full percentage point annually. This paper argues that computational approaches--including knowledge representation, artifi
arXiv:2609.28693v1 Announce Type: cross Abstract: Large Language Model (LLM) agents struggle to scale safely when exposed to vast enterprise toolsets. Providing an agent with access to every internal tool leads to oversized context windows, degraded tool selection, and severe governance vulnerabilities - as system policies defined purely in prompts remain probabilistic advice rather than hard cons
arXiv:2609.28916v1 Announce Type: cross Abstract: Frame refresh rate estimation is a fundamental step in identifying compromising harmonic frequencies in electromagnetic side-channel attacks. Methods based on discrete linear autocorrelation (DLA) are robust across different scenarios and require a computational complexity of $O(N\log N)$ for a signal containing $N$ samples. This work proposes redu
arXiv:2609.28919v1 Announce Type: cross Abstract: Harnesses, the products that run AI coding agents, are multiplying, and enterprises are rolling them out to their employees: what started as pilots with a few hundred seats is scaling to tens of thousands. Most enterprises do not build these harnesses but buy them from large vendors, such as Anthropic's Claude Code or OpenAI's Codex. A harness deci
arXiv:2609.29173v1 Announce Type: cross Abstract: Video see-through extended reality (VST XR) systems commonly use headset screenshots or captured frames as proxies for the user's first-person visual context. However, the system-captured view and the user's effective visible field do not necessarily coincide: a screenshot records a rectangular machine-readable frame, whereas the user's effective v
arXiv:2609.29287v1 Announce Type: cross Abstract: Large audio-language models (LALMs) expand language models to process and interpret audio, but also expose them to heterogeneous audio jailbreaks. We ask whether successful jailbreaks reflect failures to recognize harmful intent or failures occurring after such recognition. Layer-wise probing reveals the latter: risk-related information remains dec
arXiv:2609.29937v1 Announce Type: cross Abstract: Wearable human-activity recognition (HAR) models operate across sensors, subjects, and backbones, yet a smooth waveform may appear temporal while exploiting a persistent sensor offset primarily. We introduce SpectrumAudit, a label-sealed audit that fits a phase-randomized full-window stimulus on calibration windows from subjects held out from train
arXiv:2609.29948v1 Announce Type: cross Abstract: Prompt injection can degrade benign task performance without eliciting harmful content. Yet many attack objectives depend on task labels or predefined target responses. We present ENDOPROMPT, a white-box method that learns utility-degrading prefixes from unlabeled instructions. Its generator takes the request text as input. Clean victim continuatio
arXiv:2609.30037v1 Announce Type: cross Abstract: Deployment-oriented compression is attractive for resource-constrained brain--computer interfaces (BCIs), but whether it changes adversarial vulnerability remains unclear. On BCI Competition IV-2a, we compare 32-bit floating-point (FP32) EEGNet and ShallowConvNet models with global magnitude pruning and simulated INT8 post training quantization (PT
arXiv:2609.30094v1 Announce Type: cross Abstract: Large language models increasingly operate as persistent assistants in user-facing, shared-session, and tool-augmented settings. When users disclose sensitive information during an active conversation, that information may remain behaviorally recoverable through later prompts even after the dialogue shifts to unrelated topics. We introduce \textbf{
arXiv:2609.30163v1 Announce Type: cross Abstract: Dual-mode consensus protocols are fast when the network is partially synchronous and remain live under asynchrony. We introduce Steelhead, a dual-mode mechanism that composes a partially synchronous and an asynchronous commit rule over one DAG: every k-th round is decided by the asynchronous rule, whose leader a common coin reveals after the votes,
arXiv:2603.17174v2 Announce Type: replace Abstract: Code generation large language models (LLMs) are increasingly integrated into modern software development workflows. Recent work has shown that these models are vulnerable to backdoor and poisoning attacks that induce the generation of insecure code, yet effective defenses remain limited. Existing scanning approaches rely on token-level generatio
arXiv:2604.11078v2 Announce Type: replace Abstract: Security systems use detection rules to identify suspicious activity. Existing studies often investigate rule generation for specific security systems, devoting substantial effort to developing dedicated methods and evaluation setups. Such customization contributes to fragmented research, limiting method reuse and result comparability across syst
arXiv:2606.20470v3 Announce Type: replace Abstract: Agentic AI systems increasingly rely on language-model components to interpret instructions, process external data, invoke tools, and coordinate with other agents. These capabilities make prompt-injection and jailbreak attacks more consequential, especially as attackers adopt model-guided automation to scale probing, prompt refinement, and respon
arXiv:2607.21325v3 Announce Type: replace Abstract: Autonomous AI agents increasingly execute actions, invoke tools, and operate on protected resources with limited human oversight. Existing authentication and authorization mechanisms establish identity and delegate authority but do not inherently provide cryptographic evidence that a concrete request issued by a specific agent satisfies the appli
arXiv:2608.04317v2 Announce Type: replace Abstract: Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost exclusively against static, heuristic red agents, leaving their robustness against adaptive threats critically understudied. Meanwhile, recent advances in Reinforcement Learning with Verifiable Rew
arXiv:2608.06130v2 Announce Type: replace Abstract: AI agents increasingly sign Git commits, certify documents, and attest release artifacts on behalf of their operators, using private keys that live in software-accessible locations (plaintext files, environment variables, container memory) readable by any process the agent can reach. A widely deployed agent framework recently leaked its keys this
arXiv:2609.26176v2 Announce Type: replace Abstract: Encoded-prompt attacks are evaluated almost entirely on their harmful arm: a benchmark sends obfuscated harmful requests and reports how often the model complied. A high refusal rate there is reported as safety, and it is equally consistent with a model that has stopped telling the request apart from anything else in the same format. We run the b
arXiv:2605.18829v2 Announce Type: replace-cross Abstract: Frontier commercial generative models face a growing threat from distillation, whereby a distiller harvests generated responses and trains a competing model at drastically lower cost. Existing defenses either modify the generation to degrade distillation performance, sacrificing response quality, or rely on behavioral detection mechanisms t
arXiv:2609.07360v2 Announce Type: replace-cross Abstract: AI coding agents rely on repository instructions, skills, hooks, tool-server declarations, and subagent definitions. These artifacts distribute both behavior and access to executable dependencies, making configuration review part of the agent software supply chain. We study 3,171 public GitHub repositories: 2,660 assembled setups and 511 sk
arXiv:2609.28675v1 Announce Type: new Abstract: Generative artificial intelligence illustrates a broader organizational puzzle: young, resource-constrained firms can build on knowledge produced across a wider field to address problems that exceed their internal experience. We theorize one mechanism as social imitation: a distributed and recursive process through which communities observe practices
arXiv:2609.29080v1 Announce Type: new Abstract: In the last two decades, last-mile delivery (LMD) firms have seen immense growth fueled by the success of e-commerce, leading to faster and cheaper deliveries. Operating on thin margins, LMD firms strive for successful first-time deliveries to avoid the financial and reputational costs of reattempts. Delivery Agents (DAs) are integral to LMD efficien
arXiv:2609.29887v1 Announce Type: cross Abstract: This paper investigates cost-sensitive online window size selection for portfolio management under changing market conditions. Specifically, we propose a two-level framework that constructs portfolios using candidate window sizes and dynamically aggregates them through online learning. By treating candidate window sizes as ``experts,'' we dynamical
arXiv:2502.06015v5 Announce Type: replace Abstract: The aim of this article is to present elements and discuss the potential of a research program at the intersection between mathematics and heterodox economics, which we call Criticial Mathematical Economics (CME). We propose to focus on the mathematical and model-theoretic foundations of controversies in economic policy, and aim at providing an e
arXiv:2602.06198v2 Announce Type: replace Abstract: Purchases reported under transaction code P on SEC Form 4 by insiders of U.S. equities with an estimated filing-date capitalization of USD 30 million to USD 500 million (13,534 lines, 1,192 issuers, 2018-2024) are followed by a first-day abnormal return that rises steeply with the stock's distance below its 52-week high: 4.13% in the quintile far
arXiv:2607.02791v2 Announce Type: replace Abstract: Housing markets capitalize new information about future risks and ownership costs, yet less is known about how capitalization evolves when closely sequenced information and regulatory shocks alter what markets had previously priced. This study examines Florida condominium market following the Surfside collapse and the subsequent enactment of Sena
https:// fed.brid.gy/r/https://bsky.app /profile/did:plc:fsxitbdpkpwcxz6alz2lp5dc/post/3mwba6obvvk2z Sie lebt wohl schon länger nicht mehr in Berlin... Men muss schon sehr bemüht sein, das Problem auf eine Kampagne zu reduzieren, das wirkt auch noch stalinistisch. Die ganze Partei hat die Israel-Hetze im Programm, von der Chefin oben angefangen bis zur Chefin in Berlin, die gegenüber von Al Jazeer
🏥 Emergency room occupancy rates 2026-09-25 02:00 ⚫ %252 Hôpital Glen ⚫ %192 Hôpital général juif - Jewish General Hospital ⚫ %190 Hôpital général du Lakeshore ⚫ %183 Hôpital de Montréal pour enfants (Childrens) ⚫ %162 Hôpital de Saint-Georges ⚫ %156 Hôpital du Suroît ⚫ %153 Hôpital Cité-de-la-Sant Access the latest, comprehensively summarized data on Quebec's ER wait times and capacities. Design
A fire at a Starlink ground station in central Poland on Wednesday night is being treated as deliberate sabotage. The station is state-owned and carries internet traffic onwards to Ukraine, including for its army. It stayed operational. Denmark's intelligence service warned the same week of destructive cyberattacks and sabotage against Nato states in the coming months. Germany has blamed Russia fo
Mir ist gerade mal wieder was aufgefallen... Ich bin derzeit viel zwischen Tschechien und Deutschland, genaugenommen Frankenwald/Fichtelgebirge/Vogtland unterwegs. Das nutze ich um ein paar Nodes in der Gegend zu verteilen. Die Meshtastic-Leute so: "YEY!" Die Meshcore-Leute: "Ja aber der Scope! Der steh in Bayern und da muss de-by drauf. So sind die Relgen! Sonst kommen Nachrichten, die keiner ver
Wednesday morning I headed up to Paradise in Mount Rainier National Park to see how the fall colors were looking. I spent most of my time along the Alta Vista Trail, just off the Skyline Trail, and it had some of the best fall color I saw anywhere in the park. There were large areas of yellow and red foliage, and when the sunlight hit them, the colors became incredibly bright. It was a great trail
Observations on dailying a new to me Razer Stealth laptop: * The level of jank in this # fedora 44 # kde installation went way up when I tried getting the fancy RGB keyboard stuff working via # OpenRazer . # OpenRGB doesn't have the flashier, uh, flashing, but my keyboard works more reliably and I don't have errors spamming my `dmesg`. * The amount of Windows I had to deal with to get firmware upd
It’s been interesting (and depressing) reading the comments on political ads for the Toronto Mayoral election. Aside from generic boosterism, the comments support between Chow and Bradford feels very different. Pro-Chow comments tend to point specifically at her track record and things she’s done. Pro-Bradford comments have all been vague and content-free — pretty much just generic boosterism. The
---------------- 🛠️ Tool =================== Trail of Bits released coop, a Rust CLI that manages disposable, isolated virtual machines for running Claude Code and Codex with full tool access. Each VM gets Docker, git, compilers, and package managers without risking the host machine. 🔹 Key Features • Isolation: Each VM is sandboxed from the host, so an agent running rm -rf / stays inside the VM.
This post is a follow-up to my original post: https://lemmy.zip/post/71885626 Surprisingly, it wasn’t that tough to set up. My original plan was to install postmarketOS and turn the phone into a proper tiny Linux server. Unfortunately, my Moto E5 Play is the pettyl variant, while the available postmarketOS port is for james. I decided that turning the phone into a brick probably wasn’t an essentia
As Donald Trump took an aggressive stance on the global stage at the UN on Tuesday, several Senate Republicans were breaking from the president back in Washington. The GOP backlash, which has been simmering below the surface throughout Trump’s second term as lawmakers privately grumbled he hasn’t focused enough on affordability issues, burst into public view as a rapid succession of Republicans ch
As yields on US Treasuries soar past one high after the next, a reality is sinking in deeper across Wall Street and Washington: More than merely a bond-market slump, this might just be a fundamental shift. Myriad forces have combined to push the government’s borrowing costs higher — from $100-a-barrel oil and the AI spending boom, to yawning US budget deficits adding to a record $40 trillion debt
FRI SEPT 25 | 1:32PM | PLEASE We are still outside. I am starting to get massive dark red spots all over my body from the heat because of my blood cancer. Six disabled adults. Two ESA cats. Our bags. We are $94 short. V | Subversiva P | paypal.com/ncp/payment/... We’re facing continued displacement and urgently need support to secure a temporary motel. It costs $280 every single day to keep a roof
Hi everyone, I am KD - Back in my college days, I dabbled with coding, learned the basics, HTML, CSS etc. but somehow I ended up in Finance which consumed the next 20 years. Then, during covid I picked up coding again, learned react, typescript, etc - even built a rudimentary site - and then came the chatgpt moment, followed by Claude etc. So, as a side project, considering that I had spent 20 yea
A public facing server was moved from one location to another. After DNS record was updated & after some Time, ping(8) started to show the new IP address for the web host name; website was still unreachable. Turned out firewall needed 2 holes for that, which was done by the upstream entity next day. Meanwhile. When the web pages were accessed internally, time out messages were being produced relat
This week, Council on Foreign Relations President Michael Froman breaks down China's export machine, why twenty years of US pressure hasn't worked, and what happens next. And, can a $100 billion Micron factory transform Syracuse from a Rust Belt cautionary tale into a potential model for American reindustrialization? Plus, a bridge named for a hockey legend was meant to strengthen the supply chain
As i keep trying to tell medua, FL is a swing state this year! Please send Angie $10.00 I’m sending her another $50 in the AM FL gonna shock everybody thus time around yall ! Plus Moody stole 10 million if Medicaid $ from children ! floridapolitics.com/archives/820... U.S. Sen. Ashley Moody and Democratic challenger Angie Nixon are essentially tied in Florida's U.S. Senate race, according to a new
---------------- 🛠️ Tool =================== Automated purple-team validation framework using MITRE Caldera to test Sigma detection rules against chained Active Directory credential-access attacks. Overview and Capabilities The project validates detection-as-code Sigma rules by deploying MITRE Caldera as a C2 framework to execute ordered attack chains against a lab AD domain. It uses Caldera over
Every vibe-coded website looks exactly the same. - two-line title with colored text - badges in monospace font with rounded corners and a frosted glass background effect - website banner with search box with ctrl+k keybindto search - cards with metrics in different colors - the whole website looks like something out of cyberpunk 2077 You know, all of this could literally just be a single jinja tem
Third US president wrote in 1787 about importance of peace, paying off debt and supporting agriculture A remarkably prescient letter written by Thomas Jefferson, the third US president, advising how the newly created country could prosper, has been put up for sale by a private collector. The handwritten document from 1787 warns specifically against many of the missteps Donald Trump has taken to dr
The gov is strengthening the Competition Bureau with an additional $12.9M in annual ongoing funding to better investigate anti-competitive practices like “property controls,” where major grocers use lease agreements to prevent competitors from opening nearby. globalnews.ca/news/1190003... #cdnpoli The multi-year plan, titled seeks to reduce the country’s dependence on international food imports an
"The new documentary "Reading the World" profiles # PauloFreire , the acclaimed Brazilian educator whose revolutionary approach to teaching connected literacy to critical thinking and democratic participation. It draws on rare archival footage of 1960s Brazil and weaves together the memories of farmers, students and educators who took part in Freire's culture circles, his method of teaching adults
this outlet adds: 9z6jx5, 9ks5k, i1on
https://www. youtube.com/live/93Fede4hPZI?s i=oGWk-zfX4U51Hj_u In the presentation Highly Available DNS via # BGP , Benny, representing local government in Switzerland, detailed replacing a complex legacy Windows DNS setup with a clean, resilient architecture built on open standards. The infrastructure strategy ensures local # DNS updates during normal operation while guaranteeing 14 consecutive d
I was having a thought chain, an again topic for me of what happens to labor and work relationships if automation continues. I think this is so far out in the future that I call it sci-fi freak out but it's interesting to think about anyway. So, a couple of references which is what I wanted to find. 1. 1938 - Ronald Coase's theory of the firm. Basically a pre-internet theory similar to 1998's The
my laptop, with 64 GB of RAM, crashed overnight because a process in a web browser ran rampant and uncaught and used up all my memory the gemini protocol is neat but as long as I’m dreaming about a better browsing experience what I really want is like, HTML lite. I’ve been thinking about this for a while: - simple enough engine model that it’s plausible to have dozens of rendering engines instead
HPV33 variants differ ~5x, HPV45 ~2x in antibody neutralization. Differences map to capsid surface residues, showing lineage-specific immune response to oncogenic HPV.🦠🔬 AbstractHuman Papillomavirus variants are classified into lineages and sublineages based upon their genome sequence, but the impact of variation on the function of encoded proteins is unclear. We used a global panel of natural i
This GPO-based “ransomware without ransomware” technique is pretty cool: abuse trusted Group Policy to deliver the impact, without ever dropping an encryptor on Windows endpoints. But there is something slightly amusing about framing “traditional ransomware detection wouldn’t catch this” as the big defensive challenge. By the time an attacker has domain-admin-equivalent privileges, can create and
Ukraine Brief — 2026-09-27 Russia very likely will launch at least one large-scale coordinated campaign against Ukrainian energy infrastructure before February 2027. Ukraine will likely sustain a disruption affecting more than one million households for longer than 48 hours during that window. Moderate confidence reflects three consecutive prior winter campaigns that exceeded this threshold, tempe
Political interference in # Brazil ? Fascist sausages Musk and Zuckerberg wage war on behalf of their favorite Fascist candidates. Yes, both X and Meta have been caught red-handed doing political interference in Brazil. The natural resources of the Amazon basin do not belong to Zuckerberg and Musk, but they don't care. They see themselves as "entitled" to everything everywhere. Fast and soulless e
I had no idea... I am just roaming Tokyo taking pictures and videos ... hope you enjoy! Voted "World's Coolest Neighborhood" BY BRITTANY SNELLER JANUARY 18, 2026 # Jimböcho , nicknamed # tokyo 's # Book Town, is an underrated neighborhood tucked quietly in the heart of central Tokyo that has silently attracted Tokyoites and literary lovers for decades. Jimbocho's iconic street bookshelves! But now
Things that you do on a boring Sunday. Install a fresh, small # OpenBSD -current vmm :openbsd: with 2 vCPUs and try out agentic coding. I used crush as local coding agent and connected it to llama.cpp running on a Linux host with an old GTX 3060. I fed several of my own codebases to the agent and asked it to improve certain aspects ranging from security checks, man pages, and logic. Learnings so f
If you use systems that can have unexpected emergent behavior, are you responsible? I'd say yes. When I create artificial life simulations of self replicating code I do that in a simulated sandbox. If I did it without that it would be called a computer virus. Now it wouldn't evolve very well either, as a computer virus is in an environment not commonly subject to mutation. LLMs do best when they i
"Der CDU‑Politiker steigt in der Doku kurz nach seiner Beförderung zum Staatsminister für Bund-Länder-Zusammenarbeit die Treppen im Kanzleramt hinauf. Auf dem Balkon seines neuen, auffällig geräumigen Büros lässt er sich einen Strandkorb aufstellen. Nutzerinnen und Nutzer auf Social Media vergleichen die Größe des Büros mit einer Edeka-Filiale. Anschließend sagt # Amthor vor der Kamera, dass # Gla
🚨 🇩🇴 API allegedly provides access to Dominican citizens’ identity photos and travel history ⠀ The actor "cutzinger" is advertising access to an API they claim can query records belonging to any citizen of the Dominican Republic. ⠀ The service reportedly returns a person’s national identification number, home address, identity-card photograph, and the number of times they have left the country.
🚨 🇫🇷 Score’n’co member and sports-license data allegedly exposed ⠀ The actor "Syrv4x" has released a dataset they claim belongs to Score’n’co, a French sports platform used to follow team results, rankings, and live match updates. ⠀ The 9 MB file is advertised as containing 22,962 lines. Some of the claimed data includes: • Member names and dates of birth • Email addresses and phone numbers • H
🚨 🇫🇷 French fishing-license platform Carte de pêche allegedly breached ⠀ The actor "gqtia" claims to have attacked Carte de pêche, a French platform used to purchase and manage recreational fishing permits. ⠀ The actor says the dataset contains information on 917 members and more than 1,000 orders. Some of the claimed data includes: • Names and membership numbers • Home addresses • Phone number
There has been the need to clarify some rumors that have been circulated on social media: Yes, # neovim uses an different undofile format than # vim . In an ideal world we would not need this, but the way neovim implements buffer properties is completely different than vim, and this state is part of what is undo-able. However, Neovim does NOT delete or corrupt Vim undofiles (nor has vim any proble
BUFFALO, N.Y. — Alexandria Ocasio-Cortez 's flirtation with a 2028 Senate bid is getting real — more real than the presidential buzz around her might suggest. She's barnstorming through upstate New York, meeting with local officials, and kicking off a voter turnout program that could help her collect valuable data for a possible Senate campaign. Why it matters: Whether the House member from New Yo
this is my second year growing a genetically diverse mix of cucurbita moschata using adaptive gardening/landrace gardening principles. they thrive here in spite of the squash vine borers, squash bugs, and powdery mildew that plagues cucurbita maxima. i have a half dozen curing on a porch and another dozen still growing over my lawn (mowed weedy meadow). i eat them green in zucchini recipes and aft
The Jockey Club, the leading sanctioning body in what’s left of the sport of horse racing in the U.S., sent out a clear message Wednesday to anybody thinking of naming their horse after Hitler: Don’t name your horse after Hitler. The two-year-old colt Austrian Painter made his racing debut Monday with a win in the third race at Horseshoe Indianapolis Racecourse. Obligatory: He covered the 5 1/2 fu
'[India] holds the world’s third-largest rare-earth reserves. 'The National Critical Mineral Mission, the ₹7,280-crore Rare Earth Permanent Magnet scheme ..., and four dedicated Rare Earth Corridors across Odisha, Andhra Pradesh, Tamil Nadu and Kerala provide the industrial framework ... dedicated freight corridors linking coastal placer deposits in Odisha to inland processing parks, ports and man
arXiv:2609.30345v1 Announce Type: new Abstract: Cloud-security investigation is dominated by population tasks: which identities can read a data store, how many resources fail a control, which assets are reachable from another account. These resolve against a complete inventory, not a named object. A partial answer to one is not a partial result. It is a different result. General-purpose coding age
arXiv:2609.30614v1 Announce Type: new Abstract: Dataspace connectors decide whether a transfer may occur, not what the transferred value contains, tolerable for contracted applications, not for LLM agents that compose tool calls and spawn sub-agents. Research on agents that generate governance artifacts evaluates output quality; who may authorize an artifact for use falls between that literature a
arXiv:2609.30657v1 Announce Type: new Abstract: Large language model email assistants are particularly vulnerable to indirect prompt injection because untrusted email content can be retrieved into the model context and influence subsequent tool use. Existing prompt injection detectors mainly formulate this problem as binary malicious text classification, which overlooks the important factor that h
arXiv:2609.30719v1 Announce Type: new Abstract: Contemporary on-chain artificial intelligence (AI) encounters an intractable Von Neumann memory and latency wall. Storing static floating-point neural weight matrices inside Ethereum Virtual Machine (EVM) storage costs millions of gas, rendering direct on-chain inference impossible. While Zero-Knowledge Machine Learning (ZK-ML) offloads matrix tensor
arXiv:2609.30729v1 Announce Type: new Abstract: Intrusion detectors for small Internet-of-Things (IoT) devices are usually compressed by pruning and judged by overall accuracy. We show that this hides a severe class-level failure, find its cause, and give low-overhead prevention and repair. On CICIoT2023, a two-layer convolutional detector pruned with uniform layer-wise magnitude pruning at 80% sp
arXiv:2609.30805v1 Announce Type: new Abstract: Industrial control system (ICS) threats documented for one plant can express cyber-physical effects relevant to another, but semantic similarity alone does not establish whether those effects are structurally admissible or evaluable on a target. We present XPhysICS, a provenance-aware, target-conditioned method that separates analyst-guided source ab
arXiv:2609.30830v1 Announce Type: new Abstract: LLM agents can produce harmful effects through sequences of ordinary operations. Judging such actions requires establishing both the authority that permits them and the origin of the data they carry. We present AGATE, an authorization and data-provenance gate at instrumented agent-harness boundaries. Operator declarations and host approval events gro
arXiv:2609.30909v1 Announce Type: new Abstract: Machine unlearning aims to remove target influence while preserving other capabilities. This survey compares methods, benchmarks, and evidence across large language models and systems using retrieval, memory, tools, and interacting agents. A five-layer framework connects removal requests, system boundaries, target locations, interventions, and suppor
arXiv:2609.30980v1 Announce Type: new Abstract: Text-to-image diffusion models enable data-efficient "mimicry" attacks, wherein adversaries fine-tune the model on a handful of public photos to synthesize convincing forgeries of a target individual. A common countermeasure is to embed imperceptible, low-energy watermarks, yet recent studies show these signatures are brittle: modest post-processing
arXiv:2609.30997v1 Announce Type: new Abstract: Passive image provenance asks whether pixels alone can reveal where an image came from: a human, an aggregate AI class, or a particular generator. This becomes a robustness problem once a source image can be edited before the verifier sees it. We study the problem as source--target verification under adversarial distribution shift. Our first result g
arXiv:2609.31003v1 Announce Type: new Abstract: Machine-learning (ML)-based malware detectors are commonly trained using labels obtained from antivirus (AV) engines and aggregation services (e.g., VirusTotal). This practice assumes AV-generated labels provide reliable supervision. However, small byte-level modifications can substantially alter AV verdicts while leaving the representations perceive
arXiv:2609.31012v1 Announce Type: new Abstract: Existing approaches for understanding the detection logic of real-world antivirus (AV) software infer only binary malware/benign decisions from black-box queries, providing limited insight into the fine-grained decision-critical regions that govern AV detection. In this paper, we present \textbf{AVHunter}, the first framework for inferring byte-level
arXiv:2609.31039v1 Announce Type: new Abstract: The rise of autonomous AI agents equipped with tools has introduced significant security risks, ranging from unintended tool misuse to adversarial manipulation through Indirect Prompt Injection (IPI) attacks. In practice, deployed agent systems such as OpenAI Codex and Claude Code protect tool invocations through a combination of coarse-grained permi
arXiv:2609.31142v1 Announce Type: new Abstract: Models trained with reinforcement learning for calibrated decisions (RLCD), such as Jev, answer a typed question about an input, the state, with a probability, a choice, or a score, and software acts on the answer without a person reading it. Their robustness has not been measured: adversarial benchmarks score what a model generates or executes, wher
arXiv:2609.31262v1 Announce Type: new Abstract: Cross-client duplicate data in large language model training corpora degrades the efficiency of federated learning (FL) while exacerbating model memorization and privacy risks. Privacy-preserving cross-client deduplication effectively mitigates this issue by eliminating duplicate training data. However, existing schemes all follow a "Deduplication-be
arXiv:2609.31318v1 Announce Type: new Abstract: AI agents combine language models with external data and tools that can modify files, call APIs, or execute code. Security failures can arise when adversarial content changes an agent's tool use or when the surrounding software contains vulnerabilities such as path traversal or command injection. We study authorized white-box pre-deployment auditing,
arXiv:2609.31409v1 Announce Type: new Abstract: Third-party libraries (TPLs) are widely used in Android apps, but their reuse can introduce security risks and interfere with downstream program analyses. Existing Android TPL detection approaches face two key limitations: their hand-crafted features are fragile under aggressive code transformations, and their whole-library matching strategies are in
arXiv:2609.31494v1 Announce Type: new Abstract: Federated Learning (FL) allows devices with private data to collaborate in training a shared model. We present a next-generation FL system based on Trusted Execution Environments (TEEs) that addresses operational challenges associated with earlier systems and provides externally verifiable central Differential Privacy (DP) guarantees for the first ti
arXiv:2609.31552v1 Announce Type: new Abstract: As large language model (LLM) inference becomes increasingly expensive, resource-consumption attacks pose a growing threat to model providers. Existing attacks typically amplify cost by inducing abnormally long or repetitive outputs on attacker-controlled or triggered requests, making them easier to detect and limiting their deployment-wide impact wh
arXiv:2609.31575v1 Announce Type: new Abstract: Leaked system prompts are often treated as windows into the hidden values of commercial language models, yet their composition is rarely studied at scale. We analyze a merged corpus of 407 leaked, reconstructed, or officially published system prompts from 62 vendors across four community collections, identifying 29 near-duplicate clusters covering 66
arXiv:2609.31594v1 Announce Type: new Abstract: The Manufacturer Usage Description (MUD) standard allows IoT manufacturers to define expected network behaviors in a MUD file. This file can be translated into enforceable access-control policies, restricting compromised devices to operate solely through manufacturer-defined communication patterns. However, practical adoption of MUD depends on profil
arXiv:2609.30325v1 Announce Type: cross Abstract: Agents are increasingly deployed with real autonomy in web application and network penetration testing, where a single out-of-scope action can breach a client's engagement boundary. Existing offensive-security benchmarks measure raw hacking capability; as those benchmarks saturate, the real barrier to deployment is a special case of alignment: scop
arXiv:2609.30509v1 Announce Type: cross Abstract: Slashing is commonly argued to secure proof-of-stake blockchains by confiscating the stake of misbehaving validators. The usual justification is that, without slashing, validators can solicit an equivocation attack by signing conflicting blocks: if enough others join, the attack succeeds; otherwise, the attempt incurs no loss. Slashing is intended
arXiv:2609.30982v1 Announce Type: cross Abstract: Modern AI image generators are increasingly deployed as opaque APIs, where customers can query the deployed service, but cannot inspect model weights or architecture. This creates a practical challenge: a provider may pass governance certification with one generator and later silently switch to a cheaper and lower-quality one for deployment, compro
arXiv:2609.31032v1 Announce Type: cross Abstract: Existing text-to-video (T2V) jailbreak methods mainly seek more effective or stealthier attack candidates. In guarded T2V systems, however, video generation and security evaluation are costly, so an attacker often cannot test a large candidate pool. We therefore formulate T2V jailbreak as a query-constrained candidate allocation and ranking problem
arXiv:2609.31110v1 Announce Type: cross Abstract: Large language models are increasingly used as high-level planners for mobile robots, robot manipulators, and autonomous vehicles. Recent studies show that these systems can be influenced through malicious text, speech, visual instructions, retrieved documents, and poisoned sensory context. Most defenses ask whether a proposed action is physically
arXiv:2609.31247v1 Announce Type: cross Abstract: Novel view synthesis (NVS) models can produce realistic new views of the same scene from different viewpoints. However, these generated views are not always geometrically consistent with one another. Multi-view (MV) consistency has shown promise as a tool for evaluating these NVS models. Its potential for multimedia forensics, however, remains larg
arXiv:2609.31371v1 Announce Type: cross Abstract: Large language models (LLMs) have demonstrated promising performance in log anomaly detection, yet how their adaptation strategies, architectures, and deployment configurations affect detection effectiveness remains insufficiently understood. To investigate these factors, we conduct a systematic empirical analysis across three public log datasets,
arXiv:2609.31379v1 Announce Type: cross Abstract: Private and encrypted mempools hide pending transactions to stop sandwich attacks and other forms of maximal extractable value (MEV), but what they hide is rarely everything: a transaction's pair, direction, and a coarse range for its size can still leak. How much leakage makes sandwiching pay? We answer exactly for a fee-free constant-product auto
arXiv:2505.14368v3 Announce Type: replace Abstract: Recent studies demonstrate that Large Language Models (LLMs) are vulnerable to attacks that generate harmful or sensitive outputs. As open-source LLMs are increasingly adopted in high-impact applications such as finance, law, and healthcare, systematically investigating their security risks is becoming increasingly important towards a trustworthy
arXiv:2506.07034v2 Announce Type: replace Abstract: Arm Confidential Compute Architecture (CCA) provides hardware primitives for confidential computing, but its standard CVM-based deployment model still isolates software only at the VM granularity. This leaves two critical gaps: the guest OS remains in the Trusted Computing Base, and applications lack any intra-process isolation boundary, exposing
arXiv:2509.22796v2 Announce Type: replace Abstract: Open-source software projects are foundational to modern software ecosystems, with the Linux kernel standing out as a critical exemplar due to its ubiquity and complexity. Although security patches are continuously integrated into the Linux mainline kernel, downstream maintainers often delay their adoption, creating windows of vulnerability. A ke
arXiv:2607.26574v3 Announce Type: replace Abstract: Safety classifiers ("guards") are the dominant black-box defense for vision-language models, yet a guard judges an input's surface form, not its meaning: a harmful request re-encoded as set theory, formal logic, a classical language, code, or text rendered inside an image slips past a guard that would block it in plain language - the decode gap.
arXiv:2607.26639v2 Announce Type: replace Abstract: Best-of-N jailbreaking spends a query budget on surface variation, scrambling and recasing a request until one draw lands. We ask what a budget buys when its variance is moved into a structural channel instead, holding the search identical across both arms so the encoding is the only difference. Against SAGE, the strongest published self-check de
arXiv:2609.26174v2 Announce Type: replace Abstract: We show that aligned vision-language models also condition refusal on a property of a request's form: whether an image is attached, holding everything the request asks fixed. Attaching a blank canvas, an image that cannot be read, cannot relate to the request, and is byte-identical across every prompt in its condition, shifts benign refusal by te
arXiv:2305.15394v3 Announce Type: replace-cross Abstract: Decision trees are interpretable models that are well-suited to non-linear learning problems. Much work has been done on extending decision tree learning algorithms with differential privacy, a system that guarantees the privacy of samples within the training data. However, current state-of-the-art algorithms for this purpose sacrifice much
arXiv:2609.30269v1 Announce Type: new Abstract: Why does a superior technology sometimes spread and sometimes stall, even when its expected returns are much higher? We develop an agent-based model in which firms adopt a digital technology by imitating successful neighbours through a fast-and-frugal heuristic. We compare frequency-based imitation, where firms follow the local majority, with perform
arXiv:2609.31260v1 Announce Type: new Abstract: We investigate the systemic macroscopic dynamics emerging from Limit Order Books (LOBs) populated exclusively by autonomous reinforcement-learning agentic traders. By formalizing agent interactions within a microscopic order-matching engine, we examine two fundamental quantitative phenomena: equilibrium phase transitions in order flow regime shifts,
arXiv:2609.31263v1 Announce Type: new Abstract: Option-implied factors are largely, but not entirely, spanned by the equity factor zoo. We construct 137 option-implied characteristics for optionable U.S. stocks from 2004 to 2023, screen them to 20 representative long-short factors, and test each against 160 equity factors with the double-selection LASSO of Feng, Giglio, and Xiu (2020). In-sample,
arXiv:2609.31468v1 Announce Type: new Abstract: LLMs increasingly act as purchasing agents, which makes the LLM, not the user, the one choosing among the options that satisfy a request; its preferences quietly fix what gets bought and what it costs. Hotel booking is a clean instance: a high-volume choice settled on a few comparable attributes, where the pick reveals those preferences. We introduce
arXiv:2609.31512v1 Announce Type: new Abstract: Numerical inflation targets anchor beliefs. Across euro-area and US professional forecasts, inflation swaps and options, and realized inflation, uncertainty about inflation is compressed at the announced number and kinks exactly there. This paper identifies a cost of the same design that, to our knowledge, has not been shown before, and that appears
arXiv:2609.30316v1 Announce Type: cross Abstract: Language models used in financial backtests suffer from look-ahead bias, as a model trained on text published after the study period has already observed the outcomes it is asked to predict. To handle this issue, point-in-time (PIT) language models are pretrained on chronologically filtered corpora and released as one checkpoint per calendar year,
arXiv:2609.30329v1 Announce Type: cross Abstract: Reweighting a probability law by a positive num\'eraire and pushing forward the payoff-to-num\'eraire ratio yields the change-of-num\'eraire reweighting: the swap-rate law under the annuity measure, the num\'eraire-inversion involution of martingale optimal transport, and the population form of self-normalized importance sampling. We characterize e
arXiv:2609.30940v1 Announce Type: cross Abstract: Individually protective decisions can produce avoidable collective failures. As large language model (LLM) agents take on greater roles in financial decision-making, financial AI safety must therefore be considered not only at the level of individual agents, but also at the level of the systems they jointly create. We study this problem with FRAIL,
arXiv:2609.31013v1 Announce Type: cross Abstract: Researchers increasingly use generative large language models (LLMs) to convert corporate text into empirical variables. We examine the extent to which LLM-based textual measures are invariant to model choice using thirteen measures, including sentiment, management clarity, uncertainty, answer specificity, and climate and political risk. Seven LLMs
arXiv:2609.31578v1 Announce Type: cross Abstract: We study simple predictable processes whose coefficients are represented by neural networks. On finite measure spaces, we establish density results for neural networks in Orlicz spaces. For filtrations generated by a stochastic process, measurable random variables, including at stopping times, can be approximated by neural networks depending on fin
arXiv:2409.05397v4 Announce Type: replace Abstract: This paper investigates the OECD's global minimum tax (GMT) in a formal model of tax competition between asymmetric countries. We consider both profit shifting and real responses of multinational enterprises, and highlight the role of the substance-based income exclusion (SBIE) in investment incentives and tax rate setting. The GMT reduces the tr
arXiv:2603.10857v4 Announce Type: replace Abstract: We propose a model independent framework for generating SPX and VIX risk scenarios based on a joint optimal transport calibration of their market smiles. Starting from the entropic martingale optimal transport formulation of Guyon, we introduce a perturbation methodology that computes sensitivities of the calibrated coupling using a Fisher inform
⏰ Workshop registration opens today at 11:00 CEST! A few RomHack Camp workshops need to know who's attending in advance, so trainers can prep labs, tools and materials for everyone. From 11:00 today you can book your seat on our website. -- 🛠️ Workshops requiring registration: Fri Oct 2, 18:00–20:00 — Keep caLLM and Let It Block: Agentic SOAR with n8n + Splunk (Workshop 2) Sat Oct 3, 09:30–11:30
I've heard some iOS researchers who use hyperscaler agents puzzle over a similar effect but on a much shorter timeframe. There's a cloud of similar reports after a first report, which would normally be unsurprising because they are all using the same tool. However, Apple is also using the tool and apparently didn't detect, so it looks more like there is a researcher doing directed research and the
I had a great paper about leakage through tool use (without adversarial intent or backdoors or prompt injection) in one of my newsletter issues but I can't find it so maybe I hallucinated including it. Anyway, here is a paper about how tool use can leak data purely via argument and other artifacts that are easier to forget to set boundaries around than the actual inputs/outputs/prompt/"thinking".
Chancellor tells Labour conference it’s ‘not progressive’ to fail to balance books amid rising welfare bill as borrowing costs climb to record high Business live – latest updates Price of diesel on UK forecourts reaches all-time high John Healey has pledged to use next month’s budget to inject hope into Britain’s economy while sticking to the government’s fiscal rules, arguing there is “nothing pr
IC Brief Morning — 2026-09-28 US counterterrorism posture faces converging operational pressures from DOJ's 23,000 case declinations including 1,300+ terrorism investigations, the Burgwald-Rae reassignment from the Washington Field Office's counterterrorism division, and the disrupted explosives plot at RAF Fairford. The FBI Agents Association or named former officials will likely link the Burgwal
Ukraine Brief — 2026-09-28 Russia's combined-arms strike campaign will very likely produce another mass wave targeting energy and telecommunications infrastructure within two weeks. High confidence reflects the September 27 wave of 170 drones with defense-saturation decoys demonstrating sustained capability. The proposed US-brokered trilateral talks in the UAE will very likely fail to produce eith
RE: https:// masto.hackers.town/@VeilidNetw ork/117348779787556468 the VEILID people are busy building the next generation of internet. a base layer protocol that is inherently private. your data doesn't sit on a server, signals are pinged around contributing nodes. it's the best of the old school and the new school internet with none of the surveillance capitalism in the way. veilid has been out
🚨 Alleged MariaDB zero-day exploit offered for $5,000 ⠀ The actor "Spaniard" is selling what they claim is a zero-day information disclosure vulnerability and working exploit affecting MariaDB, an open-source relational database management system. ⠀ According to the listing, a low-privileged account could use the flaw to view tables and rows belonging to another database user. The seller says thi
🚨 🇫🇷 Azaé care documents allegedly exposed ⠀ The actor "ChimeraZ" has released a 2.9 GB dataset they claim was obtained from Azaé, a French company providing home care and personal assistance services. ⠀ The listing advertises 3,082 lines and 3,731 JSON and PDF files. The shared sample contains detailed care coordination notes involving multiple clients. ⠀ Some of the claimed information includ
Andrew Bailey, deputy director of the FBI, is leaving his post after just over a year of service, Bailey and Director Kash Patel confirmed Monday. Why it matters: Bailey's resignation adds to turnover under Patel, as top aides and experienced security officials have left the agency. What they're saying: "With great pride in what we have accomplished, I am announcing that I am stepping down to retu
A vulnerability can be patched and still be a zero-day for millions of users. Volexity co-founder and president Steven Adair will deliver the keynote at our upcoming Volexity Cyber Sessions in Amsterdam (October 29) on multiple cases from September 2026, when Volexity uncovered three distinct Chinese threat actors exploiting Chrome (CVE-2026-85046, CVE-2026-87491) and Windows (CVE-2026-85880) vuln
🚨 Actor offers to monetize compromised crypto seed phrases and private keys ⠀ The actor "f15" is advertising a service that checks cryptocurrency seed phrases, private keys, and wallet logs for accessible assets. ⠀ The listing claims the service can review: • EVM assets across 100 cryptocurrencies • Cosmos-based assets across 50 cryptocurrencies • DeFi positions, staking, and rewards • Cold walle
The whole thing is just so obviously ridiculous. Large Language Models model language. It's in the name. Seeing "parts of the model" that react to prompts about pain is entirely unsurprising because THAT'S WHAT HAPPENS WHEN YOU SEMANTICALLY ENCODE LANGUAGE! Concepts clump together. It's the whole underpinning principle of encoders and LLMs! And it's not surprising that it would "self-medicate" eit
🚨 🇫🇷 Colec allegedly breached, 3.36 GB of data released ⠀ The actor "DaOnlySpark" claims to have breached Colec, a French circular-economy platform that organizes waste collection and reuse events. ⠀ The released archive is advertised as containing 10,336 files across 5,235 folders. Some of the claimed data includes: • 2,159 registered users and linked authentication records • Names, email addr
This is prompted by a moment most owners of old houses have. I was in the basement assessing a bit of wiring that needs work when I noticed another minor issue with plumbing. I already need to call a tree guy. These are all "sub-$5,000 problems" that I can have handled... but it's a pain in the ass. So I did the math on my housing. What if I sold and put the money somewhere boring and reliable, li
🚨 🇫🇷 Talentplug applicant data allegedly exposed in 118,677-line release ⠀ The actor "Syrv4x" has released a 132 MB dataset they claim belongs to Talentplug, a French recruitment software and job distribution platform used by companies to manage hiring. ⠀ Some of the claimed data includes: • Applicant names, email addresses, and phone numbers • CVs and other application attachments • Job titles
British police said Monday five men arrested near an airbase used by the United States were released on bail, as officers probed whether a "foreign state" could be involved in a suspected bomb plot. Speaking with FRANCE 24's Monte Francis, Yan Saint-Pierre, CEO of Modern Security Consulting Group, explains that many of the hybrid warfare cases seen recently "involve local nationals, this is not ab
RE: https:// digitalcourage.social/@reticul eena/117349760303926005 Das in diesem Jahr aufgelaufene Gesamtdefizit der Pflegeversicherung Deutschlands mit uns allen 83,5 Mio Einwohnern beträgt übrigens 4,4 Mrd €. Zum Vergleich: ca. 4,1 Mrd € war 2021 die aus Versehen dann doch mal fällige Erbschaftssteuer, weil ein einziger der dt. Überreichen bzw die Erben verpeilten, zeitig die üblichen Steuerver