Google says AI agent turned patches into exploits with no human help
Original title: Google: AI agent turned a security patch into a working exploit with no human help
GTIG says a suspected Russian-speaking operator’s exploitagent framework then stole data from victim organizations.
Google Threat Intelligence Group says a suspected Russian-speaking operator used an AI agent to turn security patches into working exploits with no human in the loop, then stole data from victim organizations. GTIG calls it a first and documents it in its fourth-quarter 2026 AI threat tracker as a step beyond Q3 in fully autonomous vulnerability discovery and exploitation with LLMs. Hunters found an exposed staging server running exploitagent, a multi-agent research and weaponization framework using Russian-language tooling and frontier and local models for continuous differential analysis of popular web application software.
Unconfirmed: Russian-speaking attribution is described as suspected, and “first such occurrence” is GTIG’s assessment.
- The framework is named exploitagent and used Russian-language tooling.
- It was set to analyze popular web application software continuously.
- GTIG contrasts the case with attacker LLM use in Q3 2026.
- Victim organizations are not named in the report summary.
Impact: Data stolen from victim organizations
- Vendors
- Malware
- exploitagent
- Organizations
- Google Threat Intelligence Group
From the original article
2 min read at databreaches.netJessica Lyons reports: A suspected Russian-speaking operator used an AI agent to turn security patches into working exploits with no human in the loop – a first such occurrence, according to the Google Threat Intelligence Group (GTIG) – before ultimately stealing data from victim organizations. Google documented this attack in its fourth-quarter 2026 AI threat...
By Dissent · databreaches.net
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.