Threat
Discovering and exploiting vulnerabilities in critical, real-world software with Claude
First reported anthropic.com
Page published
Earliest dated coverage: 29 Oct 2026 · First observed: 29 Oct 2026 · Latest dated coverage: 29 Oct 2026
Coverage timeline
Single-source research — one report is available.
Why it matters
Claude’s demonstrated ability to build multi-step exploits with minimal human input highlights the potential for AI-assisted exploitation to compress defenders’ patching windows.
Anthropic reports that an unreleased Claude model built five working exploit chains targeting Chrome, Safari, Android, Windows, and Linux with little or no human interaction. Demonstrated impacts included a Chrome sandbox escape, Android data theft, and zero-click Linux root code execution; according to the coverage, all affected vendors have patched the vulnerabilities.