Release a patch and there is already a framework reading it. According to Google Threat Intelligence Group, bad actors use autonomous LLM frameworks that continuously scan software updates, figure out which vulnerability was patched and assemble an attack for it. No human involved.
There is no CVE number: this is an assessment of how attacks are built, not a flaw in a specific product.
Fact
The facts: on 16 November 2026 Google published an announcement on its blog about the latest AI Threat Tracker report from Google Threat Intelligence Group. According to Google, bad actors now use autonomous LLM frameworks to continuously scan software updates, figure out which vulnerability was patched and build an attack for it, with no human involvement. Google is also announcing Google AI Threat Defense, which the company says combines Gemini and other frontier models, Wiz's context and prioritization, CodeMender's automated patching and Mandiant's expertise to continuously find and fix flaws before attackers can exploit them. Google writes that defending against machine-speed threats at human speed is not feasible.
John Hultquist, Chief Analyst at GTIG, puts it this way: for the most part, attackers are not using new, groundbreaking methods to break into systems. They are still going after known, unfixed issues, just with much greater speed and scale. Instead of relying on manual effort to keep up, we will have to embrace agentic defense. Or we risk falling behind.
The attack no longer waits for a human to write it.
The assessment is Google's, and Google is also announcing the product it says solves the problem. The announcement is short and points to the full report; it gives no numbers on the scale of autonomous attacks.
Look at how long it takes from patch to response on your side. It is now measured against a machine that does not get tired.
Official primary sources
Google - The era of autonomous attacks has arrived, 16.11.2026The visual is generated code-art, no third-party images.