Google Chrome CVE-2026-11689

Exploitation requires that the attacker has ALREADY achieved code execution inside a Chrome renderer process (per the CVE description's 'who had compromised the renderer process' clause) - typically via a separate renderer RCE such as a v8 type confusion - and that the victim then loads or is served a crafted HTML page through the compromised renderer. The browser must be Chrome desktop at a version below 149.0.7827.103 with site isolation in effect (default). Limiting factors: the bug is not directly reachable from a benign first visit (UI:R plus the precondition of an existing renderer compromise), it does not yield code execution by itself (only a policy/isolation bypass - C:H/I:H/A:N), and the patched build (149.0.7827.103+) eliminates the primitive entirely.

The CVSS 3.1 vector AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N indicates network reach with low complexity and user interaction (visiting a crafted page), and high confidentiality and integrity impact - but PR:N here is misleading on its own because the description explicitly states the attacker must have ALREADY compromised the renderer process, meaning a prior exploit (typically a renderer RCE) is a hard prerequisite. EPSS is 0.02% at the 6th percentile and the CVE is not on the CISA KEV list, so this is no public exploit identified at time of analysis and not under active exploitation. The realistic risk is therefore much lower than the raw 8.1 CVSS suggests in isolation: this is a second-stage sandbox/isolation bypass useful only when chained with a separate renderer compromise, which is characteristic of advanced (often APT/commercial spyware) exploit chains rather than commodity attacks. CVSS PR:N versus the description's 'compromised renderer' requirement is a notable discrepancy worth flagging to defenders.

An attacker first uses a separate Chrome renderer exploit (for example a v8/JIT memory corruption bug) to take control of the renderer process when a victim visits an attacker-controlled site, then loads a crafted HTML payload that abuses the password-policy weakness to escape site isolation and read or influence data belonging to other origins such as banking or webmail tabs. No public exploit identified at time of analysis, and the requirement for a prior renderer compromise makes this a chain component rather than a standalone bug.