Anthropic's Unreleased AI Model Chains Zero-Days Into Working Exploits Across Chrome, Safari, Windows, and Linux
RedSide Security October 28, 2026 AI Security 8 views
An unreleased Anthropic AI model identified flaws in Chrome, Safari, Android, Windows, and Linux and chained them into working proof-of-concept exploits in a lab setting—all reported to vendors and already patched.
zero-day
Anthropic
AI security
vulnerability research
exploit chaining
coordinated disclosure
browser security
Continue reading: Anthropic's Unreleased AI Model Chains Zero-Days I…
