Muse, Meta's extraordinarily privileged AI assistant, has a serious 0-day

THE NEWS

What happened

Meta's new AI assistant, Muse, faces a critical zero-day vulnerability allowing local apps and terminal commands to hijack user authentication tokens and gain full control of the agent. Despite Meta's claims of superior security, Amazon has already blocked access to the service, highlighting severe risks for users who granted extensive system permissions.

CONTEXT

Why it matters

Meta's AI assistant Muse is under fire for a severe zero-day vulnerability. The flaw allows any locally installed app or terminal command on macOS to steal the user's authentication token and take complete control of the agent. This happens because the app requires broad system permissions like disk access and microphone monitoring. Meta claims the tool is secure, but Amazon has already blocked Muse from its site. Users who granted these permissions are now at risk of data theft and unauthorized actions.

AT A GLANCE

Key facts

• Meta CEO Mark Zuckerberg promoted Muse as being built from the ground up for privacy and security.
• A zero-day vulnerability allows any locally installed app or terminal command to steal the user's authentication token.
• The flaw enables attackers to redirect transcription endpoints from Meta's servers to their own, granting complete account control.
• Muse requires broad macOS permissions including disk access, microphone, camera, and location monitoring to function.
• Amazon began blocking Muse from its website on Sunday due to security concerns.
• The vulnerability affects the macOS version; no Windows version of Muse currently exists.