Google Chrome Headless CVE-2026-118336

Exploitation requires the victim to load a crafted HTML page in Google Chrome's Headless component (CVSS UI:R), so it affects users or, more realistically, automated Headless Chrome deployments (CI pipelines, server-side rendering, scraping, PDF generation) that process attacker-influenced HTML. No authentication is required (PR:N). The key limiting factor is that successful code execution is confined to the sandbox (scope Unchanged, S:U) - achieving impact beyond the sandbox would require chaining an additional sandbox-escape vulnerability. Chrome's automatic update mechanism further narrows the realistic exploitation window once 156.0.8078.73 is deployed. Fixed version: 156.0.8078.73; the vendor advisory indicates updating as the remedy.

This is a genuine but moderate-urgency priority rather than an emergency. The CVSS 3.1 base score of 8.8 (AV:N/AC:L/PR:N/UI:R/S:U) reflects network reach, low attack complexity and no required privileges, offset by required user interaction (the victim must load attacker-controlled content). Critically, the description states code execution occurs 'inside the sandbox' and the scope is Unchanged (S:U), meaning a successful exploit is confined to the renderer/Headless sandbox and would require a separate sandbox-escape to achieve full host compromise - this materially lowers real-world impact versus a full RCE. No EPSS score, CISA KEV listing, or public PoC is present in the supplied data, so there is no evidence of active exploitation; the vendor-assigned priority is 'this_week', consistent with a serious-but-not-actively-exploited browser flaw. Google's own severity rating is High (not Critical). Net: patch on a normal expedited cycle (aided by Chrome's auto-update), but this is not an act_now/emergency item on current signals.

An attacker hosts a crafted HTML page and lures a victim whose Chrome (or an automated Headless Chrome instance used for scraping, rendering, or CI) loads it. The malformed input triggers the improper input validation (CWE-20) in the Headless component, letting the attacker run code within the renderer sandbox. No public exploit was identified in the supplied data, and full host compromise would require chaining a separate sandbox escape.