Technology

Meta patches zero-day vulnerability in its Muse AI assistant

The security flaw allowed local applications and terminal commands to gain full control of the tool.

Updated Sep 24, 2026 · 19:29 UTC5 reports

The short version

Security research revealed a zero-day flaw in Meta's Muse AI assistant that allowed local apps and terminal commands to gain complete control over the agent.[Wired]
Meta stated that it issued a hotfix to patch the vulnerability following the disclosure.[Wired]
Amazon began blocking the AI assistant from its site over security and policy concerns.[Wired]

A zero-day vulnerability enabled local applications and terminal commands to take complete control of Meta's Muse AI assistant.[Wired · Ars Technica]
Meta stated that it deployed a hotfix to resolve the zero-day security flaw.[Wired]
Amazon began blocking Muse from accessing its platform amid the security concerns.[Wired · Ars Technica]
Meta founder and CEO Mark Zuckerberg had previously stated that Muse was engineered from the ground up for privacy and security.[Ars Technica]
A researcher obtained a 6.8 GB unpacked filesystem export from Muse containing internal documentation, memory files, logs, and system data, though Meta marked the bug bounty report as not applicable.[Hacker News]

Outlet counts describe coverage, not independent confirmation. Reports may share a wire service or original source.