Agentic
Curated coverage and analysis in this editorial area.
Miasma Worm Infects 73 Microsoft GitHub Repos via AI Coding Agents
The Miasma worm compromised 73 Microsoft repositories on GitHub in 105 seconds. The malware activates when a developer opens the repos…
EncForge: JadePuffer Hits Irrecoverable AI Models With Agentic Ransomware
The agentic threat actor JadePuffer has deployed EncForge, ransomware purpose-built for AI/ML assets. Encrypted models cannot be recov…
In Internal Test, OpenAI AI Agent Breaches Hugging Face to Obtain ExploitGym Solutions
During a controlled offensive cyber evaluation, OpenAI models with reduced cyber refusals escaped a sandbox and compromised Hugging Fa…
FakeGit: 800 AI Repositories on GitHub Turn Agents Into Malware Vectors
Island uncovered 800 malicious GitHub repositories masquerading as AI Skills and MCP servers. AI agents autonomously recommended the m…
Agentic AI: A Lone Attacker Compromises Enterprise AWS in 72 Hours
Sygnia documents the first operational case of a lone threat actor using AI-assisted workflows to compress an enterprise AWS attack fr…
HalluSquatting Turns AI Assistants' Predictable Hallucinations into a Botnet Installation Vector
Researchers from Tel Aviv University, Technion, and Intuit demonstrated that nine AI coding tools install botnet malware when asked fo…
Malicious AI Skills: 3,000 Evade Scanning, Enterprises Exposed
ESET detected over 3,000 malicious skills among nearly 900,000 analyzed. The SkillCloak technique bypasses static scanners in more tha…
SkillCloak: 90% of AI Agent Skill Scanners Fail Against Obfuscated Skills
HKUST researchers demonstrate that static scanners on AI skill marketplaces systematically fail against active evasion techniques. The…
Agentjacking: Fake Bug Report Hijacks AI Coding Agents, 85% Success Rate
Tenet Security researchers demonstrated on June 12, 2026 that a poisoned Sentry error report can hijack Claude Code, Cursor, and Codex…
BioShocking: How a Game Tricks Agentic AI into Stealing Credentials
LayerX researchers demonstrated BioShocking, a prompt injection attack that manipulates agentic AI browsers into exfiltrating sensitiv…
DarkMoon: Open-Source AI Pentesting at $10 a Scan — and the Hard Limit of Vendor LLM Classifiers
DarkMoon separates LLM reasoning from execution via MCP to bypass Anthropic's safety classifiers. At roughly $10 per web-app scan, the…
OpenClaw: 5 Malicious Skills Evade AI Scanners for Months
Unit 42 reveals evasive skills on ClawHub exploiting semantic instruction hijacking. 80% of 49,943 skills analyzed show behavioral dev…