Start your day with intelligence. Get The OODA Daily Pulse.
Google Threat Intelligence Group (GTIG), security researchers observed the first documented instance of malicious actors utilizing a fully autonomous AI framework to convert public software security patches into working exploit code without any human-in-the-loop oversight. GTIG notes that rather than expending high computational effort on traditional zero-day discovery, threat actors are increasingly weaponizing large language model agentic loops against n-day vulnerabilities, leveraging vendor patch data as finite, high-probability source material to reverse-engineer exploits at machine speed. The discovery highlights a major evolution in cyber threat tactics, shifting from AI-assisted coding to end-to-end automated exploit pipelines where autonomous agents handle vulnerability analysis, trial-and-error debugging, and final attack generation. Cybersecurity experts emphasize that the milestone drastically compresses the traditional “patch window,” the critical time organizations have to apply updates before attackers can weaponize the underlying flaw, requiring automated defense-in-depth and accelerated patch management across global enterprise infrastructure.
Full report : Threat actors deploy agentic LLM workflows to analyze vendor patches and generate attack code with zero human intervention.