Google Chrome CVE-2026-17784

Exploitation requires two specific preconditions: (1) the attacker must have already compromised the Chrome renderer process through a separate vulnerability - this is a chained attack, not a standalone exploit; and (2) the victim must be running Google Chrome on macOS at a version prior to 151.0.7922.72 and interact with a crafted HTML page (UI:R). The vulnerability is limited to the macOS platform; Chrome on Windows or Linux is not indicated as affected. What limits exploitation: the mandatory prior renderer compromise is a high bar requiring an additional, separate exploitation chain, making this predominantly a threat from well-resourced actors rather than opportunistic attackers. The PR:N in the NVD CVSS vector reflects the absence of authenticated privileges required to deliver the initial page, but does not account for the renderer pre-compromise condition.

The provided CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) yields 8.8, which overstates realistic single-step risk - the description explicitly states exploitation requires a prior renderer compromise, which is a significant additional prerequisite not captured by AC:L or PR:N alone. The scope metric S:U also appears inconsistent: a successful sandbox escape, by definition, changes the exploitable scope beyond the renderer process. In practice, this is a chained vulnerability, meaning real-world exploitability depends on pairing it with a separate renderer exploitation bug. EPSS at 0.21% (11th percentile) aligns with this assessment: low observed exploitation probability, consistent with a complex multi-stage attack rather than a trivially weaponizable flaw. CISA KEV status is not confirmed, and no POC is publicly known. The Chromium team rated severity as Medium internally, which better reflects the preconditions. Priority should be moderate-high for macOS Chrome deployments, primarily for environments targeted by sophisticated threat actors capable of chaining multiple renderer bugs.

A threat actor targeting a macOS user first exploits a separate renderer vulnerability (via a different bug) to gain code execution within Chrome's sandboxed renderer process. The attacker then serves a crafted HTML page that triggers the Audio UAF, corrupting heap memory in the renderer to manipulate the sandbox escape path and achieve arbitrary code execution on the underlying macOS host. No public proof-of-concept exists at time of analysis, making this attack chain primarily a risk from sophisticated adversaries capable of independently developing or chaining browser exploits.