CVE-2026-5907
vulnerability analysis and mitigation

Overview

CVE-2026-5907 is an out-of-bounds memory read vulnerability in the Media component of Google Chrome, caused by insufficient data validation when processing crafted video files. It was reported by Luke Francis on February 15, 2026, and publicly disclosed on April 7–8, 2026, as part of the Chrome 147 stable channel release. The vulnerability affects all Google Chrome versions prior to 147.0.7727.55 on Windows, Mac, and Linux, as well as Microsoft Edge (Chromium-based). It carries a CVSS v3.1 base score of 8.1 (High), though Google internally rated it Low severity (Chrome Releases, GitHub Advisory).

Technical details

The root cause is classified as CWE-125 (Out-of-bounds Read): Chrome's Media component fails to adequately validate data boundaries when parsing video file content, allowing a read operation to access memory beyond the intended buffer. An attacker exploits this by delivering a specially crafted video file to a victim — for example, via a malicious web page or a downloaded file — which Chrome's Media subsystem then processes. No elevated privileges are required, but user interaction (opening or navigating to content that triggers video processing) is necessary. The Chromium issue tracker references bug ID 484665123, though full technical details remain restricted pending broad user adoption of the patch (Chrome Releases, GitHub Advisory).

Impact

Successful exploitation can result in high confidentiality impact — potentially exposing sensitive data from Chrome's process memory — and high availability impact, which may include application crashes or instability. Because the vulnerability is an out-of-bounds read rather than a write primitive, direct code execution is unlikely without chaining additional vulnerabilities; however, memory disclosure could facilitate information leakage or serve as a stepping stone in a multi-stage attack. The scope is limited to the Chrome renderer process and does not directly affect other system components (GitHub Advisory, Chrome Releases).

Exploitation steps

  1. Craft a malicious video file: Prepare a specially crafted video file (e.g., MP4, WebM, or another format supported by Chrome's Media component) that contains malformed metadata or stream data designed to trigger an out-of-bounds read in Chrome's media parsing code.
  2. Deliver the payload: Host the malicious video file on an attacker-controlled web page, or distribute it via email, messaging, or file-sharing platforms to lure the victim into opening it with a vulnerable Chrome version (prior to 147.0.7727.55).
  3. Trigger processing: When the victim visits the page or opens the file in Chrome, the browser's Media component processes the crafted video, triggering the insufficient data validation flaw and causing an out-of-bounds memory read.
  4. Harvest leaked memory data: Depending on the memory layout, the out-of-bounds read may expose adjacent memory contents — potentially including sensitive data such as tokens, credentials, or other in-memory values — which could be exfiltrated back to the attacker via JavaScript or network requests.
  5. Chain for further impact (optional): Use the disclosed memory information (e.g., heap addresses) to defeat ASLR and chain with a separate memory corruption vulnerability for potential code execution (Chrome Releases, GitHub Advisory).

Indicators of compromise

  • Network: Unexpected outbound connections from the Chrome process to unknown external IPs shortly after a user opens a video file or visits a media-heavy page; HTTP requests delivering unusually structured video files (e.g., malformed MP4/WebM with anomalous container metadata).
  • Logs: Browser crash reports or minidumps referencing Chrome's Media component or media pipeline stack frames; repeated renderer process crashes logged in Chrome's internal crash reporter.
  • Process: Chrome renderer processes (chrome.exe / chrome on Linux/Mac) crashing or restarting unexpectedly after loading video content; unusual child process behavior associated with media decoding.
  • File System: Suspicious video files downloaded to the user's Downloads directory from untrusted or unknown sources, particularly with unusual file sizes or metadata inconsistencies.

Mitigation and workarounds

The primary remediation is to update Google Chrome to version 147.0.7727.55 or later (147.0.7727.55/56 on Windows/Mac, 147.0.7727.55 on Linux), which contains the fix for this vulnerability (Chrome Releases). Microsoft Edge (Chromium-based) users should also apply the corresponding Edge update. As a temporary workaround prior to patching, organizations should advise users to avoid opening video files from untrusted sources and to be cautious when visiting unfamiliar websites that may auto-play video content. Enterprise administrators should enforce automatic Chrome updates via policy to minimize exposure windows.

Community reactions

The Chrome 147 release was covered by security-focused outlets including GBHackers, which highlighted the broader batch of 60+ vulnerabilities patched in this update, including two critical WebML flaws (GBHackers). BeyondMachines also noted the significance of the Chrome 147 update in the context of the critical WebML vulnerabilities. CVE-2026-5907 itself, rated Low severity by Google internally, received limited individual attention given its lower risk profile relative to the critical and high-severity issues in the same release. No notable researcher commentary specific to this CVE has been identified.

Additional resources


SourceThis report was generated using AI

Free Vulnerability Assessment

Benchmark your Cloud Security Posture

Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.

Request assessment

Get a personalized demo

Ready to see Wiz in action?

"Best User Experience I have ever seen, provides full visibility to cloud workloads."
David EstlickCISO
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
Adam FletcherChief Security Officer
"We know that if Wiz identifies something as critical, it actually is."
Greg PoniatowskiHead of Threat and Vulnerability Management