Executive brief
A vulnerability exists in Google Chrome's AI component that could allow an attacker to bypass security protections. If an attacker has already compromised the part of the browser that displays web pages, they could use this flaw to escape the 'sandbox'—a security layer designed to prevent malicious websites from accessing the rest of your computer. This could lead to unauthorized access to your files or system settings.
Technical details
An improper input validation vulnerability (CWE-20) exists in the AI component of Google Chrome prior to version 151.0.7922.72. The flaw allows a remote attacker who has already achieved code execution within the sandboxed renderer process to bypass sandbox restrictions. By providing specially crafted input to the AI component via a malicious HTML page, the attacker can potentially gain elevated privileges on the host system. This vulnerability is mitigated by the requirement of a prior renderer compromise. Google has addressed this issue in the stable channel update for desktop.
Affected products
- Google Chrome prior to 151.0.7922.72
Timeline
- 2026-07-29: patched: Fixed in version 151.0.7922.72
- 2026-07-30: disclosed