Meta patches Muse exploit that let attackers control the AI agent

<figure> <img alt="Robot hands typing on a keyboard" data-caption="The zero-day exploit required local access to the user’s device, but gave potential attackers access to Muse accounts. | Image: The Verge" data-portal-copyright="Image: The Verge" data-has-syndication-rights="1" src="https://platform.theverge.com/wp-content/uploads/sites/2/2026/09/STKS533_AI_AGENTS_HACKING_D_54a015.png?quality=90&#038;strip=all&#038;crop=0,0,100,100" /> <figcaption> The zero-day exploit required local access to the user’s device, but gave potential attackers access to Muse accounts. | Image: The Verge </figcaption> </figure> <p class="wp-block-paragraph">Meta has issued a patch for its Muse macOS app following the discovery of a zero-day vulnerability that could allow someone to take control of the AI agent. The bug found by <a href="https://x.com/patrickwardle/status/2102045926474785265">security researcher Patrick Wardle</a> utilized an undocumented Muse setting that enabled potential attackers running local code to redirect transcription processing from Meta's servers to their own endpoint, <a href="https://arstechnica.com/security/2026/09/muse-metas-extraordinarily-privileged-ai-assistant-has-a-serious-0-day/"><em>Ars Technica</em></a> reports, giving the attacker access to the Muse account.</p> <p class="wp-block-paragraph">Several design decisions reportedly enabled this flaw, including having Muse dictation occur in the cloud instead of on-device, and allowing any app to control all of Muse's undocumented settings. Pr …</p> <p><a href="https://www.theverge.com/tech/998679/meta-muse-patch-zero-day-exploit-ai-agent">Read the full story at The Verge.</a></p>