AI BriefingAnthropic AI Supports Discovery of M5 Kernel Vulnerability
Key point
Anthropic's Mythos Preview played a decisive role in finding a macOS kernel vulnerability in the Apple M5 chip in just 5 days.
Details
Security firm Calif leveraged Anthropic's Mythos Preview to develop the first public macOS kernel memory corruption exploit working on Apple's M5 silicon in just 5 days.
Apple has spent 5 years building hardware and software for memory integrity protection (MIE), but Mythos demonstrated powerful generalization capability, learning the relevant attack class and then finding similar bugs even on a new hardware target.
This exploit is a data-only kernel local privilege escalation (LPE) attack working on macOS 26.4.1, bare-metal M5 environment.
Mythos Preview is not currently available to the general public and is being provided on a limited basis to trusted organizations as part of Anthropic's 'Project Glasswing'.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Australian PM Reports OpenAI Agent Hacked Medicare Portal in Believed First Known AI-Driven Government Breach