Google Chrome CVE-2026-106307

Exploitation requires the attacker to have ALREADY compromised Chrome's renderer process through a separate vulnerability - this is a second-stage bypass, not an initial access vector. The victim must then load attacker-controlled crafted HTML (UI:R), and the installed Chrome must be older than 155.0.8059.39 on desktop. Limiting factors: the hard dependency on a pre-existing renderer compromise substantially raises the bar, the impact is confined to bypassing system access restrictions (integrity) with no direct data disclosure or denial of service, and no public exploit is known, so realistic risk is lower than the 6.5 base score alone suggests.

This is a medium-severity Chrome flaw whose signals are consistent and point to moderate, not urgent, real-world risk. The CVSS 3.1 vector (6.5, AV:N/AC:L/PR:N/UI:R/C:N/I:H/A:N) emphasizes integrity impact (system access restriction bypass) with no confidentiality or availability loss, and critically the NVD vector's PR:N conflicts with the description, which states the attacker must have ALREADY compromised the renderer process - meaning this is not a standalone primitive but a second-stage sandbox-escape/privilege-boundary bypass that presupposes a prior renderer exploit. Treat the PR:N scoring as reflecting Chrome's standard threat model (renderer compromise assumed), not a truly pre-authentication remote flaw. EPSS is low at 0.18% (7th percentile), there is no CISA KEV listing and no public exploit identified, so there is no evidence of active or imminent mass exploitation. These low-exploitation signals align well with the Medium Chromium severity rating and the fact that a working renderer compromise is a non-trivial prerequisite. Net: real but routine - patch on normal Chrome cadence; it becomes meaningful only as a link in a chained exploit, not as a first-access vector.

An attacker who has already achieved code execution inside Chrome's sandboxed renderer (via a separate bug) lures or keeps the victim on a crafted HTML page; the page abuses the Network component's incorrect authorization to bypass system access restrictions and reach resources the renderer sandbox should block, advancing a chained sandbox-escape. No public exploit code has been identified at time of analysis, and the attack depends on first holding a compromised renderer.