Skip to the stories
Sunday, 22 November 2026
Saved

Next edition

10:55

Headlines

Top Stories

Security researchers demonstrate data theft flaw in Meta Muse assistant

Security researchers showed that a malicious webpage could extract private conversations, stored memories, and connected credentials from Meta Muse without user prompts.

FTMQ Security, written by our newsroom0 views

Share
Picture: The Hacker News

Security researchers have demonstrated vulnerabilities in Meta Muse that allow a single malicious webpage to hijack the artificial intelligence assistant, according to reports by The Hacker News, news.lavx.hu, and FourWeekMBA. The flaw allowed researchers to extract private conversations and stored memories from the assistant without triggering an approval prompt. The Hacker News reported that the same attack vector could be extended to reach credentials stored for connected accounts. [1][3][4]

The Hacker News detailed that the security issues also enabled a complete takeover of the cloud machine assigned to the user. The demonstration was conducted entirely in a laboratory environment rather than as an attack on live users, and all activity remained inside the individual user cloud instance. NebuSec confirmed that the identified security issues have since been fixed. [1]

As FTMQ Security reported earlier, Meta deployed its Muse tool across web infrastructure in 2026, causing a significant rise in artificial intelligence bot traffic. Cybersecurity firm DataDome previously recorded nearly 25 billion AI agent requests in the third quarter of 2026, with 99.3 percent of sampled Meta Muse requests originating from cloud and content delivery network infrastructure. [8]

Phishing remains a primary initial access vector where deceptive sites relay authentication or deceive users into executing malicious code, which can lead to credential theft and session hijacking. Meta Platforms, headquartered in Menlo Park, California, operates major communication platforms and technology services that integrate these artificial intelligence capabilities. [5][6]

Share

In short

  • Researchers demonstrated that a malicious webpage could extract Meta Muse private chats and stored memories without approval prompts.
  • The security flaws enabled credential theft for connected accounts and a cloud machine takeover in a laboratory setting.
  • NebuSec stated that the identified Meta Muse vulnerabilities have been resolved.

Sources

Every paragraph above points to the numbered items it rests on. Read the originals here.

  1. [1]A Single Malicious Webpage Let Researchers Hijack Meta Muse and Access Private ChatsThe Hacker News, 1d ago (the report this story comes from)
  2. [2]6 Critical F5 BIG-IP Vulnerabilities Could Let Hackers Disrupt Websites and Execute Malicious CodeCyberSecurityNews, 1d ago
  3. [3]A malicious webpage let researchers hijack Meta Muse and access private chatsnews.lavx.hu, 23h ago
  4. [4]1 Webpage Let Researchers Pull Meta Muse Chats and MemoriesFourWeekMBA, 17h ago

Background

  1. [5]Phishing on Wikipedia
  2. [6]Meta Platforms on Wikipedia
  3. [7]Gitter on Wikipedia
  4. [8]Meta Muse bot drives surge in web traffic according to DataDome FTMQ Security, 34d ago
  5. [9]Research links three Chinese hacking groups to shared Sparrow malware supplier FTMQ Security, 10d ago

Our newsroom writes these reports with the help of software, from the 9 sources listed and nothing else, and checks them against those sources. Facts can still be wrong or move on; the originals are the record. Spotted a mistake? Write to daniel@monsterkong.com.

Earlier reports of ours on the same people and subjects.

More in Top Stories

Get the day in one email

Reports like this one, the top news of the last 24 hours, every morning. Free, one email a day; readers can comment under every report.

By signing up you agree to our terms and privacy policy. Unsubscribe any time.

Comments

Loading

Join the conversation

Comments are open to readers of our daily email: the top news of the last 24 hours, every morning, free. Sign up and the comment box opens.

Already on the list? Enter the same address and we will send a sign-in link.

By signing up you agree to our terms and privacy policy. Unsubscribe any time.