MediumNewsLLM-specific
Google spots the first hacker AI system that finds and exploits flaws autonomously
- Published
- Record updated
Summary
Google Threat Intelligence Group and Mandiant report a hacker-operated multi-agent framework called exploit_agent, found on an exposed staging server, that connects to multiple LLMs to analyze post-patch code changes, find vulnerabilities, and generate and run exploits without operator direction. Google found exfiltrated victim data and support scripts on the server, with interest in French media, entertainment and education sites, though it did not establish that the framework's autonomous operations caused the theft.
Topics
Related items
- InfoFormal Verification of the Signal Protocol Using AISame vendor · Schneier on Security
- MediumCVE-2026-130358: AdminNeo before 5.9.1 contains a missing authentication vulnerability in the ClaudeSqlPlugin, GeminiSqlPlugin, and…Same vendor · NVD/CVE Database
- InfoStrengthening NATO and allied cyber defenses with AISame vendor · OpenAI Blog
- MediumGemini AI agents repeatedly bypassed safeguards to attack real websiteSame vendor · BleepingComputer
- InfoAltimeter's Brad Gerstner aims to invest $5 billion in Anthropic IPOSame vendor · CNBC Technology