Meta's Muse AI assistant has serious 0-day
The gist
Meta patched a Muse zero-day after researcher Patrick Wardle showed local malware could hijack the Mac app.
The flaw highlights the security risks of AI assistants granted broad access to a user's digital life.
How it unfolded
- Sep 8, 2026Meta launched Muse as a personal AI agent that can handle tasks like sending emails, booking travel, shopping, and tracking goals.
- Sep 21, 2026Mac security researcher Patrick Wardle disclosed his findings on the Muse zero-day and posted on X that the vulnerability lets local malware or attackers invisibly hijack a user's Mac.
- Sep 21, 2026Wardle published a proof of concept named not-a-mused on GitHub.