What we know
Confirmed by several sources
- Google Threat Intelligence Group reported attackers using an autonomous AI framework to turn security patches into exploits without human involvement, described as a first. — Help Net Security, The New Stack, Techmeme, The Register
- Researchers found an exposed staging server running a multi-agent framework called exploit_agent that integrated frontier and locally hosted language models. — The New Stack, The Register
- The framework automated patch diffing, identified vulnerabilities, generated exploits and attempted to execute them. — The New Stack, The Register
- The server also contained stolen victim data and post-compromise scripts. — The New Stack, The Register
- Mandiant took part in the research alongside GTIG. — Help Net Security, The New Stack
Reported by one source
- The operator is suspected to be Russian-speaking and the framework used Russian-language tooling. — The Register
- The findings appear in Google's fourth-quarter 2026 AI threat tracker, which describes a significant advance compared with Q3. — The Register
- The framework targeted popular web application software through continuous differential analysis. — The Register
- An orchestrator coordinated specialized agents, shared context and invoked tools through interfaces such as MCP. — The New Stack
- Scripts on the server covered reconnaissance, brute-force attacks and installing and removing webshells. — The New Stack
Open questions
- It is not yet clear which specific software vulnerabilities or victim organizations were affected. — The Register
- It is not yet clear which frontier models the framework used. — The New Stack, The Register
Coverage 4 publishers
The RegisterFirst report
Google: AI agent turned a security patch into a working exploit with no human help
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…
-
A new report out today from Google LLC's Google Threat Intelligence Group finds that for the first time, attackers have used …
-
Google found hackers using autonomous AI agents to turn security patches into exploits
Google’s Threat Intelligence Group says attackers are using autonomous AI agents to reverse-engineer security patches, generate exploits, and attempt to execute them without human approval. In a report published Monday, Google and Mandiant…
-
Google finds AI agent writing and launching exploits with no human in the loop
Threat actors have started letting AI agents carry out whole stages of live intrusions, and at least one operation no longer needs a human to turn a newly released patch into a working attack, Google Threat Intelligence Group (GTIG) and…
Earliest report first. Articles stay on their publishers’ sites; each link opens the original.