Marimo CVE-2026-75149: High-Severity MCP Injection Fixed 🔗 https:// cybersecurefox.com/en/marimo-c ve-2026-75149-mcp-command-injection # Marimo # CVE -2026-75149 # MCP # notebook # vulnerability # co
mastodon:infosec-exchange 29d ago
Multiple unrelated cybersecurity incidents reported: Claude Code targeted by symlink import exploits and fake installers delivering MacSync infostealer; critical Gitea remote code execution vulnerability disclosed; NYC Health + Hospitals reports 11TB data breach claimed by LeakNet group. Anthropic also confirmed worldwide Claude service outage.
Six unrelated security stories: VMware ESXi critical patch, AI-enhanced phone fraud, US humanoid robot ban, MCP bridge vulnerability, Analog Devices data breach, and Microsoft phishing attacks. No single coherent narrative connects these items.
Eight unrelated items spanning April 2024 to November 2026: WHO partnership meetings with Netherlands and Spain, a research paper on LLM agent security, a weather alert, congressional bill tracking, FCC rulemaking, and an AI labor-market report. No coherent single story.
A critical vulnerability in Ruflo called CVE-2026-59726 (RufRoot) exposes 233 MCP tools without authentication, enabling remote code execution, LLM API key theft, and persistent memory poisoning. The flaw survives patching attempts, posing sustained security risks.
A CVSS 10.0 vulnerability dubbed RufRoot was discovered in Ruflo, an AI platform bridge tool, allowing unauthenticated attackers to execute commands and access AI provider API keys, stored conversations, and persistent agent memory. The flaw was a complete authentication bypass.
1) lets admins run arbitrary host commands. 4 now.
CISA added CVE-2026-85046, a type confusion vulnerability in Google Chromium V8 that allows remote code execution via crafted HTML, to its Known Exploited Vulnerabilities catalog as of September 4, 2026. The catalog simultaneously tracked four other critical flaws in LiteLLM, Starlette, Kestra, and SonicWall appliances.
Cluster mixes Fairphone Linux phone charging, Middle East defense pact, DOE nuclear projects, and RDP honeypot data with no single story. Reports span consumer hardware, international security, energy policy, and cybersecurity intelligence.
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 substantially more tokens than needed for normal task execution.
On 27–28 August 2026, OpenAI led an open letter signed by nearly 130 organizations—including Anthropic, Microsoft, Google, Amazon, and Cisco—warning of escalating AI-enabled cyberattacks and urging coordinated global cyber defenses. The coalition cautioned that sophisticated AI-powered attacks could proliferate within months without coordinated action.
Cluster mixes an Iran military statement with unrelated WHO announcements about Ebola response in DRC, newborn screening, and pandemic agreement negotiations. These are clearly separate stories with no connection.
Noma Labs disclosed CVE-2026-59726, a maximum-severity vulnerability (CVSS 10.0) in Ruflo, an open-source AI orchestration harness used with Claude Code and OpenAI Codex. Unauthenticated attackers can execute arbitrary commands and poison AI agent memory.
Disparate reports cover Samsung Galaxy S27 camera redesigns (2026-08-25), Anthropic's Claude memory integration (2026-08-25), Bangkok flooding and California ocean phenomena (2026-09-28), and an NRC workforce assessment (2026-09-29). A systemd SSD issue also circulated online (2026-08-27).