Google Chrome CVE-2026-17930

Exploitation requires that the attacker has ALREADY compromised the renderer process - this is the explicit, stated prerequisite and the primary limiting factor; the bug is not exploitable directly from an unmodified renderer. The attacker must then get the compromised renderer to load a crafted HTML page (UI:R - user interaction such as visiting a page is involved) that triggers the extensions input-validation flaw to perform privilege escalation. Additional limiting factors: high attack complexity (AC:H), affects only Chrome desktop below 151.0.7922.72, and Chromium's own Low severity rating indicates the impact is confined to escaping the renderer sandbox rather than immediate host compromise. This is not a no-conditions default-config remote exploit despite the PR:N metric.

The signals are mixed and should be read carefully. The supplied CVSS 3.1 vector (AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H = 7.5 High) reflects high impact, but its PR:N understates the real precondition: the description explicitly requires an attacker to have ALREADY compromised the renderer process, which is a substantial prerequisite captured only partially by AC:H. Chromium's own severity rating is Low, contradicting the 7.5 numeric score - this divergence indicates a chained, sandbox-internal privilege escalation rather than a standalone remotely-triggerable RCE. There is no KEV listing and no public exploit identified at time of analysis, and no EPSS score was provided, so exploitation probability data is missing. Net assessment: this is a defense-in-depth / chain-completion bug that matters mainly as the second stage of a multi-bug exploit; it is a routine-priority browser patch, not an emergency, but should still be remediated on the normal Chrome update cadence.

An attacker first exploits a separate renderer vulnerability to gain code execution inside Chrome's sandboxed renderer, then navigates that compromised renderer to a crafted HTML page that abuses the extensions input-validation gap to escalate privileges out of the renderer context. Given AC:H and the renderer-compromise prerequisite this is a chained, targeted attack rather than a point-and-click drive-by; no public exploit is identified at time of analysis.