nest-admin: attackers can forge admin login tokens
A secret key used to sign login tokens was accidentally saved in a file that shipped with the app. Anyone who finds it can create fake admin logins and take full control.
| Severity | CriticalCVSS 9.3 of 10 |
|---|---|
| Being exploited? | UnclearNo reports either way yet |
| Fix available? | Yesnest-admin 1.0.1 · since 27 Oct 2026 |
| How sure is this? | Official sourceBacked by an official source |
How the attack works simplified
- 1Attacker finds secretSecret is committed in .env.production
- 2Forges tokenSigns token with the known JWT_SECRET
- 3Sends token to appToken looks valid to JwtStrategy
- 4Gains admin roleRbacGuard accepts claimed admin role
- ResultAttacker gets full admin access
What happened
nest-admin versions before 1.0.1 ship a hardcoded JWT_SECRET in .env.production, which JwtStrategy and RbacGuard use to validate JSON Web Tokens. Attackers can sign their own tokens with this known secret and impersonate any user or claim the admin role.
Are you affected?
You are affected if you run nest-admin before 1.0.1 with the JWT_SECRET from .env.production in use.
nest-admin versions before 1.0.1
What to do
- Update nest-admin to version 1.0.1 or later
- Generate a new, random JWT_SECRET and set it in your environment, not in committed files
- Remove .env.production from the repository and check git history for the leaked secret
- Invalidate any existing tokens and review admin accounts for unexpected changes
Timeline
- 27 Oct 2026, 14:16 UTCPublicly disclosed
- 27 Oct 2026, 14:16 UTCReported by NVD · NVD
- 27 Oct 2026, 14:19 UTCPicked up by Zero Radar
Sources 1 report, oldest first
Official vendor, government CERT or advisory database
- CVE-2026-117564NVD · 27 Oct 2026, 14:16 UTC
Technical details
- ID
- CVE-2026-117564
- CVSS
- 9.3 / 10
- EPSS
- 0.6% chance of exploitation within 30 days
- Disclosed
- 27 Oct 2026, 14:16 UTC
- On Zero Radar
- 27 Oct 2026, 14:19 UTC (3 min after disclosure)
- Last change
- 27 Oct 2026, 14:19 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