Google GTIG says attackers used 'exploit_agent' to fully automate vulnerability discovery and exploitation
Google's Threat Intelligence Group (GTIG) reports it observed a server running an AI framework called "exploit_agent" that links to multiple language models and appears to compare code changes, identify vulnerabilities, create exploits, and execute them with no human in the loop. Configuration files include Russian-language comments and server data indicate targets in French media, entertainment, and education; GTIG says these attacks currently trade stealth for speed, leaving forensic traces.