What is the CVSS score of CVE-2026-10925?

CVE-2026-10925 has a CVSS 3.1 base score of 8.3 (High). CVSS vector: CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H. EPSS exploitation probability: 0.1%.

Which versions of Google Chrome are affected by CVE-2026-10925?

Google Chrome on macOS versions before 149.0.7827.53 contains a sandbox escape vulnerability (CVSS 8.3) that allows attackers to gain complete control over the browser and access sensitive…. A patch is available.

How to fix or mitigate CVE-2026-10925?

24 hours: Alert security teams and macOS administrators; enable automatic Chrome updates or prepare Chrome 149.0.7827.53 or later for deployment. 7 days: Deploy Chrome 149.0.7827.53 or later to all macOS systems using software deployment tools or verify auto-update completion. 30 days: Confirm 100% patching compliance across all macOS Chrome installations and implement mandatory Chrome auto-update policies.

Google Chrome CVE-2026-10925

EUVD-2026-34374 HIGH

Out-of-bounds Write (CWE-787)

2026-06-04 chrome-cve-admin@google.com

GHSA-jpc8-5g78-vg34

Google Memory Corruption Buffer Overflow Red Hat Suse Chrome

8.3

CVSS 3.1 · NVD

Severity by source

NVD PRIMARY

8.3 HIGH

AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H

SUSE

HIGH

qualitative

Red Hat

8.2 HIGH

qualitative

Primary rating from NVD.

CVSS VectorNVD

CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H

Attack Vector

Network

Attack Complexity

High

Privileges Required

None

User Interaction

Required

Scope

Changed

Confidentiality

High

Integrity

High

Availability

High

Lifecycle Timeline

Analysis Generated

Jun 05, 2026 - 02:39 vuln.today

CVSS changed

Jun 05, 2026 - 02:22 NVD

8.3 (HIGH)

CVE Published

Jun 04, 2026 - 23:16 nvd

HIGH 8.3

CVE Published

Jun 04, 2026 - 23:16 nvd

UNKNOWN (no severity yet)

DescriptionCVE.org

Out of bounds write in Skia in Google Chrome on Mac prior to 149.0.7827.53 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)

AnalysisAI

Sandbox escape in Google Chrome on macOS prior to 149.0.7827.53 lets a remote attacker who has already compromised the renderer process break out of the Chromium sandbox via a crafted HTML page that triggers an out-of-bounds write in Skia. Google rates the underlying Chromium issue High severity and a vendor patch is available, though no public exploit code has been identified at time of analysis.

Technical ContextAI

Skia is the 2D graphics library that Chromium uses to rasterize HTML, CSS, and Canvas content; on macOS portions of Skia run inside higher-privileged GPU/browser processes that are reachable from the renderer through Mojo IPC. CWE-787 (Out-of-bounds Write) indicates the flaw writes attacker-controlled data past an intended buffer boundary inside Skia, which on a process with broader macOS entitlements than the sandboxed renderer can be leveraged to corrupt memory and pivot out of the sandbox. The bug is macOS-specific per the description, suggesting the affected code path or memory layout is unique to the Mac build of Chrome/Skia rather than a cross-platform Skia defect.

RemediationAI

Vendor-released patch: update Google Chrome on macOS to 149.0.7827.53 or later via the Stable channel as described in the Chrome Releases advisory (https://chromereleases.googleblog.com/2026/06/stable-channel-update-for-desktop.html); the underlying Chromium tracker is https://issues.chromium.org/issues/500071763. Because the bug requires a pre-existing renderer compromise, interim compensating controls include enforcing Chrome auto-update via MDM, deploying Site Isolation and the macOS-specific 'enable-features=MacSyscallSandbox' hardening already on by default, and restricting high-risk browsing through a separate browser profile or per-site renderer isolation policy; trade-offs are minimal beyond the brief restart required for the update. Operators of Chromium-derived browsers should pull the corresponding Skia fix once their upstream merges land and re-image any macOS endpoint where a renderer compromise is already suspected, since sandbox escape implies the host may already be fully attacker-controlled.