CVE-2026-10960: Chrome Sandbox Escape via Uninitialized Codec Variable

Technical summary

This vulnerability stems from improper use of an uninitialized variable (CWE-457) within Chrome's codec implementation. When Chrome's renderer process—which runs in a restricted sandbox—attempts to decode a maliciously crafted media stream embedded in an HTML page, the uninitialized data leads to a predictable or exploitable state that permits sandbox escape. The attacker must first establish renderer-process compromise, typically through another vulnerability or social engineering, then trigger the codec flaw to elevate privileges and execute arbitrary code with OS-level permissions. The vulnerability affects Chrome versions prior to 149.0.7827.53 and impacts all major operating systems where Chrome runs.

Business impact

Exploitation enables complete system compromise following initial renderer compromise. An attacker can steal sensitive data, install persistent malware, pivot to other systems on the network, or establish a backdoor. For enterprises, this expands the blast radius of browser-based attacks: a vulnerability in one web application or malicious website could cascade into infrastructure-wide compromise if users visit the site. Browser users relying on sandbox isolation for protection when visiting untrusted sites face heightened risk. This vulnerability essentially neutralizes one of Chrome's key defense-in-depth mechanisms.

Affected systems

Google Chrome prior to version 149.0.7827.53 is directly affected on Windows, macOS, and Linux. The vulnerability also lists Linux kernel and macOS in affected vendors_products, suggesting the exploitation path may involve OS-specific bypass techniques or that those platforms' kernel or OS-level protections may be circumvented during the sandbox escape. Organizations running Chrome on any of these operating systems should prioritize updates.

Exploitability

Exploitability requires a two-stage attack: first, the renderer process must be compromised through a separate vector (another Chrome vulnerability, drive-by download, or user-assisted malware). Second, the attacker must trigger the codec flaw via a crafted HTML page. The CVSS vector (AC:H, UI:R) reflects this complexity—attack complexity is high and user interaction is required. However, the impact is severe (C:H, I:H, A:H, S:C with scope change). This is not currently listed on the KEV catalog, meaning active exploitation in the wild has not yet been widely documented, though the high Chromium severity rating suggests Google treats it as a priority threat.

Remediation

Update Google Chrome to version 149.0.7827.53 or later. Chrome's auto-update mechanism will typically deploy this patch automatically, but verify completion in Chrome's About menu (chrome://about). For enterprise deployments, use Chrome Enterprise policy to enforce timely updates. No workarounds exist; patching is the sole mitigation. Consider deploying additional sandboxing at the OS level (e.g., Windows Sandbox, KVM, or container isolation) for high-risk browsing workloads if immediate patching is delayed.

Patch guidance

Deploy Chrome version 149.0.7827.53 or later across all user endpoints and devices. For Windows and macOS, enable automatic updates if not already configured. In corporate environments, push the patch via Google Chrome Enterprise or your endpoint management system. Linux users should update Chrome via their distribution's package manager or directly from Google's repositories. Verify patch application by navigating to chrome://version and confirming the version number matches or exceeds 149.0.7827.53. Test critical web applications post-patch to ensure compatibility. Prioritize vulnerable instances handling sensitive workloads or exposed to untrusted web content.

Detection guidance

Monitor for unexpected sandbox escapes or privilege escalation events originating from Chrome processes, particularly following crash dumps or abnormal termination. EDR solutions should flag codec-related crashes followed by system-level process creation from Chrome's process tree. Network detection is limited; this flaw is triggered locally. Log review should focus on renderer process crashes and subsequent lateral movement indicators. Once patched, monitor continued attempts to trigger the codec path as reconnaissance. Threat hunting should correlate Chrome crashes with subsequent elevated process spawning to identify potential exploitation attempts that occurred before patching.

Why prioritize this

Although not yet in the KEV catalog, this vulnerability merits immediate patching due to its high CVSS score (8.3), scope change (S:C means system-wide impact), and severe confidentiality, integrity, and availability impact. Sandbox escape vulnerabilities are strategically valuable to attackers and often motivate coordinated exploitation campaigns. The prerequisite renderer compromise lowers urgency slightly—it is not a single-click exploit—but the sandbox escape itself represents a critical capability gap. Any organization where Chrome security is foundational to defense strategy should treat this as urgent. The combination of high impact and Chromium's designation as

Risk score, explained

The CVSS 3.1 score of 8.3 reflects high impact (all confidentiality, integrity, and availability dimensions), but mitigated by high attack complexity and user interaction requirements. The attacker cannot exploit this in isolation; they need prior renderer compromise and must craft a specific HTML page. However, the scope change (S:C) is critical—it means exploitation breaks the sandbox boundary and affects the entire system, not just the browser. The score appropriately reflects a severe vulnerability that requires chaining but has catastrophic consequences. Compared to simpler web-based attacks, this warrants higher prioritization due to its system-level impact once triggered.

Frequently asked questions