Meta reportedly patches Muse AI agent hijacking flaw
The exploit required local access, The Verge reports. Cybersecurity News says malware already running under a user account could use the macOS flaw to intercept dictated prompts and steal authentication material.
TLDR
The Verge reports that Meta patched an exploit that let attackers control its Muse AI agent, with local access required. Cybersecurity News describes a zero-day vulnerability in the macOS agent that could allow malware already running under a user account to hijack the assistant, intercept dictated prompts, inject malicious instructions and steal authentication material.
Combined views
—
3 Sources, first seen ago
