danmu_api lets anyone change server settings when default token is used
If this app runs with its default login token and no admin token set, anyone on the internet can change its settings, lock out the real operators, and point the server at other systems.
| Severity | CriticalCVSS 9.3 of 10 |
|---|---|
| Being exploited? | UnclearNo reports either way yet |
| Fix available? | Yesdanmu_api 1.21.7 or later · since 26 Oct 2026 |
| How sure is this? | Official sourceBacked by an official source |
How the attack works simplified
- 1Unauthenticated attackerNo login needed
- 2POST to /api/env/setDefault TOKEN accepted
- 3Sets ADMIN_TOKENOperators are locked out
- 4Changes OTHER_SERVER/PROXY_URLEnables SSRF
- 5Triggers redeploymentConfig changes take effect
- ResultAttacker controls server configuration
What happened
An authentication bypass in danmu_api allows unauthenticated POST requests to /api/env/set, /api/env/add, and /api/env/del to modify environment configuration when the default TOKEN is in use and ADMIN_TOKEN is unset.
Are you affected?
You are affected if you run danmu_api versions before 1.21.7 using the default TOKEN with ADMIN_TOKEN unset.
danmu_api before 1.21.7 (default TOKEN in use, ADMIN_TOKEN unset)
What to do
- Update danmu_api to version 1.21.7 or later
- Set a strong, unique ADMIN_TOKEN and TOKEN value in the environment
- Check the environment configuration for unexpected changes such as a new ADMIN_TOKEN, OTHER_SERVER, or PROXY_URL
Timeline
- 26 Oct 2026, 14:16 UTCPublicly disclosed
- 26 Oct 2026, 14:16 UTCReported by NVD · NVD
- 26 Oct 2026, 14:21 UTCPicked up by Zero Radar
Sources 1 report, oldest first
Official vendor, government CERT or advisory database
- CVE-2026-116485NVD · 26 Oct 2026, 14:16 UTC
Technical details
- ID
- CVE-2026-116485
- CVSS
- 9.3 / 10
- EPSS
- 2.3% chance of exploitation within 30 days
- Disclosed
- 26 Oct 2026, 14:16 UTC
- On Zero Radar
- 26 Oct 2026, 14:21 UTC (4 min after disclosure)
- Last change
- 26 Oct 2026, 14:21 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