Search This Blog

Wednesday, November 18, 2026

Google Says Hackers Now Build Attacks From Patches Autonomously

 Google Threat Intelligence Group says attackers are using autonomous AI frameworks to scan software patches, identify the flaw each update fixes and build an exploit before companies can install it.



Google Threat Intelligence Group and Mandiant said in a Nov. 16 report that adversaries have reached “a critical inflection point,” shifting from basic prompting to autonomous multi-agent frameworks.

“In the past quarter, GTIG identified threat actors using LLM frameworks to fully automate vulnerability discovery and exploitation without a human in the loop for the first time,” the report said.

A companion Google post said “bad actors are now using autonomous LLM frameworks to continuously scan software updates, figure out what flaw was patched, and build an attack for it — all without human involvement.”

“Defending against machine-speed threats at human speed isn’t feasible,” it added.

Researchers Find an Automated Pipeline

GTIG said it found an exposed staging server running a multi-agent framework it dubbed exploit_agent.

The tool “compares code before and after patch announcements, analyzes findings to identify potential vulnerabilities, drafts exploits, and automatically executes them without human intervention.”

Researchers also found stolen victim data and scripts for reconnaissance, brute-forcing, installing webshells and removing other attackers’ webshells.

Victim telemetry “suggests the operator was specifically interested in targeting French websites associated with the media and entertainment industries, as well as the education sector.” GTIG did not name the operator, and Newsweek reported that Google also did not say how many organizations were compromised.

Open-Source Tools in Live Intrusions

GTIG said threat actors are also using open-source frameworks such as ARTEX, CAIRN, CyberStrikeAI and PentAGI in live operations.

It observed CAIRN against targets including in the U.S. aerospace sector. From late August to early October 2026, a customized PentAGI platform was used to scan organizations across Asia.

The report said the methods “remain the same,” including compromised credentials and unpatched systems, but “the speed and scale at which frameworks can exploit these known weaknesses introduces a new defensive challenge.”

The warning tracks an earlier assessment that AI cyber threats could outpace defenses in months, and reporting on autonomous agents used against Taiwanese government networks.

A Shrinking Window for Defenders

GTIG found that disclosures and exploits nearly doubled between January and August 2026, while zero-days rose from a 2025 monthly average of eight to 11, then spiked to 22 in August.

The group said the frameworks “currently trade stealth for velocity,” leaving forensic traces, and that some LLM-generated exploits “are claimed to work but don’t.”

“As the sophistication of these tools and underlying LLMs increases, we anticipate not only reduced human involvement in these operations but also increased stealth and sophistication of the activity,” it said.

Google said it is answering with Google AI Threat Defense, which uses Gemini and other models, Wiz, CodeMender and Mandiant “to continuously find and fix flaws before attackers can exploit them.”

https://clashreport.com/world/articles/google-says-hackers-now-build-attacks-from-patches-autonomously-oopcrzv100d

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.