CVE-2026-12009: Chrome Sandbox Escape on macOS – Patch Guidance

Technical summary

CVE-2026-12009 is a sandbox escape vulnerability in Google Chrome's Accessibility subsystem on macOS. The vulnerability arises from inadequate input validation (CWE-20) that permits a remote, network-based attacker to craft malicious HTML that, when processed by an already-compromised renderer process, can break out of Chrome's sandbox isolation. The attacker must have existing code execution in the renderer context; the vulnerability then provides a mechanism to elevate privileges or escape confinement. Chrome's sandbox is designed to limit damage from renderer compromises, so its circumvention represents a critical elevation in attack capability.

Business impact

A successful sandbox escape could allow an attacker to access system resources, credentials, or user data on the host macOS system beyond what the sandboxed browser process normally permits. For organizations, this increases the blast radius of a renderer compromise from 'contained malware in browser' to 'potential system-level compromise.' Combined with user-focused social engineering (malicious websites), this creates a plausible attack chain against employees on macOS, potentially leading to credential theft, lateral movement, or data exfiltration.

Affected systems

Google Chrome versions prior to 149.0.7827.115 running on Apple macOS are affected. Windows and Linux versions of Chrome are not mentioned in the advisory. Organizations deploying Chrome on macOS should prioritize patching.

Exploitability

Exploitation requires two conditions: first, the attacker must already have achieved code execution within Chrome's renderer process (via a separate vulnerability, malware, or user compromise); second, the user must visit a attacker-controlled website. The attack is not remotely triggerable from an uncompromised state, but the combination of renderer vulnerabilities and this sandbox escape significantly lowers the bar for effective system compromise. CVSS scoring reflects the high attack complexity and the requirement for user interaction, but the consequences are severe.

Remediation

Update Google Chrome on macOS to version 149.0.7827.115 or later. Automatic updates are enabled by default in Chrome; verify that your fleet has completed the update. For organizations with auto-update disabled, manually push this patch across all macOS endpoints running Chrome.

Patch guidance

Google Chrome's built-in auto-update mechanism should deliver version 149.0.7827.115 automatically. Verify completion by navigating to chrome://settings/help, which displays the current version and prompts update if needed. For managed Chrome deployments (via MDM or Group Policy), ensure your enterprise policies allow or enforce this patch version. No workarounds are available; patching is mandatory.

Detection guidance

Monitor Chrome process behavior on macOS for signs of sandbox escape attempts or suspicious system calls from the Chrome process. Endpoint Detection and Response (EDR) tools should flag attempts to access system resources (e.g., file reads from outside normal user directories, system library injection, or privilege escalation APIs) originating from Chrome. Additionally, log failed browser updates or version mismatches to identify systems not yet patched. Network-based detection is limited since exploitation requires local renderer compromise; focus on host-level telemetry.

Why prioritize this

This is a HIGH-severity vulnerability (CVSS 8.3) with a critical severity tag from Google. Although it requires pre-existing renderer compromise, sandbox escapes fundamentally break Chrome's multi-process security model and substantially increase the value of any renderer vulnerability. For macOS deployments, this should be prioritized for immediate patching given the high consequences and the relative ease of chaining with other browser exploits.

Risk score, explained

The CVSS 3.1 score of 8.3 (HIGH) reflects: (1) high impact across confidentiality, integrity, and availability; (2) requirement for user interaction (UI:R); (3) high attack complexity (AC:H) due to the need for pre-existing renderer compromise; (4) changed scope (S:C), indicating the vulnerability can affect resources beyond the renderer. The score does not downgrade further because sandbox escapes, even when conditional, represent severe privilege escalation within a multi-process browser.

Frequently asked questions