Google Chrome CVE-2026-17745

Two conditions must both be satisfied: first, the attacker must have already achieved compromise of the Chrome renderer process through a separate, independent vulnerability - this is the primary gating factor and is reflected in AC:H. Second, the victim must actively browse to an attacker-controlled or compromised web page containing the crafted HTML content (UI:R). The vulnerability does not require any authenticated session to any service (PR:N per CVSS), but the renderer pre-compromise is a non-trivial prerequisite that significantly limits the realistic attacker population to those capable of chaining multiple browser exploits. No specific Chrome feature flag, non-default configuration, or platform restriction is described - all Chrome installations prior to 151.0.7922.72 on any desktop OS are in scope once the renderer pre-condition is met.

The CVSS 5.8 Medium score and AC:H vector jointly reflect that this vulnerability is not autonomously exploitable - it is the second stage of a chained attack requiring a prior renderer compromise. EPSS at 0.21% (11th percentile) corroborates low observed exploitation pressure. No CISA KEV listing and no public POC have been identified, placing this in the 'no public exploit' category. Despite the relatively moderate CVSS score, sandbox escapes represent disproportionately high strategic value for advanced threat actors running multi-stage browser exploitation chains; nation-state and sophisticated criminal actors routinely chain renderer bugs with sandbox escapes. The C:L/I:L/A:L ratings may understate post-escape impact since actual system-level damage depends on the attacker's subsequent payload - these metrics reflect the OOB read's direct contribution, not the full chain impact. The primary risk-reduction signal is AC:H: organizations that keep Chrome updated eliminate this link in a potential chain even if a renderer zero-day exists.

An attacker who has independently exploited a renderer-level vulnerability in Chrome - for example, through a separate JavaScript engine or HTML parser bug - then delivers a crafted HTML page that triggers the Skia out-of-bounds read, leveraging the memory disclosure or corruption to escape the renderer sandbox. Once outside the sandbox, the attacker can execute code at the privilege level of the Chrome OS process, enabling credential theft, persistence, or lateral movement. No public POC has been identified at time of analysis, and the chained nature of the attack requires substantial exploit development capability.