MuseBot admin login can be forged with a built-in default key
Anyone who knows a secret password that ships with MuseBot can fake a login as the administrator, without needing a password. From there they can read private data and shut down the bots.
| Severity | CriticalCVSS 9.3 of 10 |
|---|---|
| Being exploited? | UnclearNo reports either way yet |
| Fix available? | YesUpgrade to MuseBot 2.0.0 or later. · since 30 Oct 2026 |
| How sure is this? | Official sourceBacked by an official source |
How the attack works simplified
- 1Attacker finds default keyKey 'muse_bot_session_key' is public
- 2Forges session cookieSigns cookie with the known key
- 3MuseBot accepts cookieTrusts the forged user_id
- 4Admin access gainedReads hashes, chats, configuration
- ResultAttacker takes over as administrator
What happened
MuseBot versions before 2.0.0 use the hard-coded default session key 'muse_bot_session_key'. Because this key is publicly known, an unauthenticated attacker can sign a forged session cookie containing a chosen user_id and gain administrator access.
Are you affected?
You are affected if you run MuseBot before version 2.0.0 and have not changed the default session_key.
MuseBot versions before 2.0.0 using the default session_key 'muse_bot_session_key'.
What to do
- Upgrade MuseBot to version 2.0.0 or later
- Check that session_key is no longer 'muse_bot_session_key' and set a long random value
- Review admin accounts and chat and configuration logs for unexpected access
- Rotate passwords and any secrets stored by MuseBot if the default key was in use
Timeline
- 30 Oct 2026, 15:16 UTCPublicly disclosed
- 30 Oct 2026, 15:16 UTCReported by NVD · NVD
- 30 Oct 2026, 15:18 UTCPicked up by Zero Radar
Sources 1 report, oldest first
Official vendor, government CERT or advisory database
- CVE-2026-119660NVD · 30 Oct 2026, 15:16 UTC
Technical details
- ID
- CVE-2026-119660
- CVSS
- 9.3 / 10
- Disclosed
- 30 Oct 2026, 15:16 UTC
- On Zero Radar
- 30 Oct 2026, 15:18 UTC (2 min after disclosure)
- Last change
- 30 Oct 2026, 15:18 UTC
The plain-language parts and the attack diagram were written by an AI model (anthropic/claude-haiku-5.5) from the sources above. Facts like versions, exploitation and dates come from the sources themselves. How Zero Radar works