Postiz: signed-in users can read server files and crash the backend
Any user with a normal account can make the server reveal its secret settings, such as the keys that protect logins and passwords for connected services. A single request can also shut the server down.
| Severity | CriticalCVSS 9.9 of 10 |
|---|---|
| Being exploited? | UnclearNo reports either way yet |
| Fix available? | UnclearNo fixed version published in the sources yet |
| How sure is this? | Official, 2 sourcesBacked by an official source |
How the attack works simplified
- 1Attacker registersSelf-registration is open by default
- 2Signed-in requestAsks backend to run internal helper function
- 3Backend reads fileChannel token set to a file path
- 4Secrets exposedEnvironment holds session signing secret and credentials
- ResultAttacker gets server secrets or crashes it
What happened
Postiz's backend does not restrict which internal functions a signed-in user can trigger on a connected channel. One such helper reads files from disk based on a stored channel token, and another makes the process exit.
Are you affected?
You are affected if you run Postiz with open self-registration (the default) or have untrusted users with accounts on your instance.
Postiz (versions not specified in the provided material)
What to do
- Check whether your Postiz instance allows self-registration and disable it if you do not need it
- Restrict account creation to trusted users until a fix is installed
- Rotate the session signing secret, database password and connected service credentials if the instance was open to untrusted users
- Watch the Postiz project and GitHub advisory GHSA-h5r5-h782-vqvp for a patched release and update as soon as one is available
Timeline
- 23 Oct 2026, 23:17 UTCPublicly disclosed
- 23 Oct 2026, 23:17 UTCReported by NVD · NVD
- 25 Oct 2026, 17:20 UTCGitHub Advisory reported on it · GitHub Advisory
Sources 2 reports, oldest first
Official vendor, government CERT or advisory database
- NVD23 Oct 2026, 23:17 UTC
- Postiz lets a signed-in user ask the backend to run a function on one of their connected channels...GitHub Advisory · 24 Oct 2026, 00:32 UTC
Technical details
- ID
- CVE-2026-115931 · GHSA-h5r5-h782-vqvp
- CVSS
- 9.9 / 10
- EPSS
- 0.5% chance of exploitation within 30 days
- Disclosed
- 23 Oct 2026, 23:17 UTC
- On Zero Radar
- before the radar started
- Last change
- 25 Oct 2026, 17:20 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