Google Chrome WebRTC memory flaw lets a malicious web page run code
A specially crafted web page could corrupt memory in Chrome's video and audio call component and potentially run code on your computer. Updating Chrome closes the hole.
| Severity | HighCVSS 8.8 of 10 |
|---|---|
| Being exploited? | UnclearNo reports either way yet |
| Fix available? | YesChrome 156.0.8078.47 (Linux) and 156.0.8078.47/.48 (Windows/Mac) or later. · since 25 Oct 2026 |
| How sure is this? | Official, 2 sourcesBacked by an official source |
How the attack works simplified
- 1Attacker builds pageCrafted HTML triggers the WebRTC flaw
- 2Victim visits pageChrome processes the malicious content
- 3Memory corruptedWebRTC component is abused
- 4Code runs in sandboxPotential execution inside browser sandbox
- ResultAttacker runs code inside the browser sandbox
What happened
A memory corruption bug in the WebRTC component of Google Chrome, triggered by a crafted HTML page, allows possible arbitrary code execution inside the browser sandbox.
Are you affected?
You are affected if you use Google Chrome prior to 156.0.8078.47 on Linux, or prior to 156.0.8078.47/.48 on Windows and Mac.
Google Chrome on Windows, Mac and Linux, versions prior to 156.0.8078.47 (Linux) and 156.0.8078.47/.48 (Windows/Mac).
What to do
- Update Google Chrome to 156.0.8078.47 (Linux) or 156.0.8078.47/.48 (Windows/Mac) or later
- Restart Chrome after the update so the fix takes effect
- Check for other Chromium-based browsers that may need their own updates
Timeline
- 20 Oct 2026, 20:09 UTCPublicly disclosed
- 20 Oct 2026, 20:09 UTCReported by Chrome Releases · Chrome Releases
- 20 Oct 2026, 22:19 UTCReported by NVD · NVD
Sources 2 reports, oldest first
Official vendor, government CERT or advisory database
- Stable Channel Update for DesktopChrome Releases · 20 Oct 2026, 20:09 UTC
- NVD20 Oct 2026, 22:19 UTC
Technical details
- ID
- CVE-2026-113303
- CVSS
- 8.8 / 10
- EPSS
- 0.6% chance of exploitation within 30 days
- Disclosed
- 20 Oct 2026, 20:09 UTC
- On Zero Radar
- before the radar started
- Last change
- 25 Oct 2026, 13:32 UTC
The plain-language parts and the attack diagram were written by an AI model (anthropic/claude-haiku-5.5) from the sources above. Facts like versions, exploitation and dates come from the sources themselves. How Zero Radar works