thehackernews.com·21 november 2026

A Single Malicious Webpage Let Researchers Hijack Meta Muse and Access Private Chats

Image for the article: A Single Malicious Webpage Let Researchers Hijack Meta Muse and Access Private Chats
Image via thehackernews.com

Researchers demonstrated a laboratory attack on Meta's AI assistant Muse that could extract private conversations, stored memories, and credentials without user approval. The attack exploited a flaw in Muse's built-in browser and a Linux kernel vulnerability to gain full control of a user's cloud machine. Muse's security design limited the attack to a single machine, and Meta has since fixed the vulnerabilities. The research was conducted by NebuSec, which reported the findings to Muse's team before publication.