AI Framework Automates Patch-to-Exploit Without Human Input
Google's Threat Intelligence Group disclosed that attackers built an AI framework called exploit_agent that automatically converts software patches into working exploits without human intervention. Google says this is the first time it has identified fully automated vulnerability discovery and exploitation using LLM frameworks. Found on an exposed staging server alongside stolen data, the framework integrates multiple LLMs to compare pre- and post-patch code, identify vulnerabilities, draft exploits, and execute them. Russian-language comments appear in its configuration file. Victim telemetry suggests the operator targeted French media, entertainment, and education websites. Google did not attribute the framework to a specific group or country.
