CVE-2026-85045
CVSS 7.5 HIGH: race condition in V8 in Google Chrome prior to 152.0.7977.82 allowed a remote attacker to execute arbitrary code inside the sandbox via a crafted HTML page. EPSS 0.3% (20th percentile). Microsoft patch: Release Notes.
Vulnerabilities · 7h ago
XBOW said it built a three-bug Google Chrome full-chain in six days using OpenAI's public GPT-5.6 Sol model, and Google confirmed and fixed the bugs while paying the team a $250,000 reward. The chain ran code on Windows from a web page alone, with no click or download required.
The exploit crossed Chrome's nested sandboxes by stitching together bugs in separate parts of the browser. XBOW said the AI searched source code, tested ideas, drafted proofs of concept, and helped assemble the chain, while humans picked the targets and combined the pieces. That cuts the time and cost to produce a working browser chain, which matters because Chrome's update only takes effect after a restart.
For teams that rely on browser isolation or fast auto-update as the main defense, the exposure sits in the gap between patch release and browser restart. If Chrome on Windows stays open, the fix is present but not yet active, and the report shows smaller teams can now reach that gap much faster.
CVSS 7.5 HIGH: race condition in V8 in Google Chrome prior to 152.0.7977.82 allowed a remote attacker to execute arbitrary code inside the sandbox via a crafted HTML page. EPSS 0.3% (20th percentile). Microsoft patch: Release Notes.
1 source covering this story
Researchers Used a Public OpenAI Model to Build a $250,000 Chrome Full-Chain Exploit in Six Days
XBOW built an AI-assisted, three-bug Chrome exploit chain in six days that enabled Windows code execution; Google has fixed all three bugs.
Part of the PlainSec briefing for 2026-10-14