Google observes autonomous multi-agent frameworks in live intrusions
Google Threat Intelligence Group and Mandiant reported multiple threat actors using open-source multi-agent offensive frameworks in live intrusions and, for the first time in their telemetry, a framework that automatically compared patches, identified vulnerabilities, drafted exploits and executed them without a human in the loop.
Why it matters: Patch disclosure can now feed automated n-day weaponization pipelines, while stolen model credentials and cloud quotas are monetized through LLMJacking. Defenders have less time between public fixes and scaled exploitation.