ZDF
Sign inSign up

Vulnerabilities

Vulnerability · GHSA-xf86-hfpm-92xr

AI Chat route exposes admin-only tools to non-admin users

Published Oct 19, 2026 · catalog updated Oct 19, 2026 · AI Chat route exposes

Last checked Nov 1, 2026. A connected source was read within the last 7 days.

Open the advisory

Exploitation likelihood is not on this record for GHSA-xf86-hfpm-92xr. A missing EPSS score is not a low risk. Compare AI Chat route exposes with the linked advisory before you change production.

Gaps on this record

  • The stored product name is a package slug. The name on this page is taken from the title (AI Chat route exposes).
  • No vendor fix or required action is stored. Compare the linked advisory before you change production.
  • EPSS is missing. A missing score is not a low risk.
Severity
Low3.1
Known exploitation
Not in CISA/VulnCheck KEV
EPSS
Awaiting synchronization with FIRST
Remediation
Version ranges on file
Affected products
npm directus

### Summary The AI Chat endpoint ( ) did not enforce the admin-only restriction on AI tools. Any authenticated user with app access could request the , , , and tools, which are reserved for administrators. The MCP endpoint enforced this restriction correctly. ### Impact The tools ran with the requesting user's own accountability, so they did not grant permissions beyond what the user already had: - , and are still guarded by the admin checks in their underlying services. - and are subject to the regular permission system. A non-admin user could only read or modify flows and operations their policies already allow, the same as through the REST API. The issue is that the admin-only tool boundary was not applied consistently. Instances where non-admin policies grant access to or are most affected. We recommend keeping those collections admin-only, as described in the [security best practices](https://directus.io/docs/configuration/security-best-practices). Exploitation requires an AI provider to be configured on the instance. ### Workarounds - Disable the AI Chat feature by removing the configured AI provider. - Make sure non-admin policies have no permissions on and .

Are you affected?

Enter the version you run. The result uses only the ranges stored on this record.

The check returns affected, not affected, or incomplete. It does not guess a version that is missing from the record.

  • npm directus>= 11.14.0, < 12.3.0

How to fix GHSA-xf86-hfpm-92xr

  • Upgrade

    GHSA-xf86-hfpm-92xr lists a fix boundary at 12.3.0. Compare the installed build with that boundary before you change production.

  • Version check

    Affected < 12.3.0

  • Advisory

    Open the inventory_github record and follow its fix.

  • Official advisories

What to do

Three questions, each answered from this record only. Confirm the vendor advisory before you patch.

Should you patch?

Exploitation likelihood is not on this record for GHSA-xf86-hfpm-92xr. A missing EPSS score is not a low risk. Compare AI Chat route exposes with the linked advisory before you change production.

Which versions are affected?

Compare what you have installed with >= 11.14.0, < 12.3.0. Compare the installed version with the linked vendor advisory before you change production.

What should an analyst verify?

CWE-863: Incorrect Authorization. EPSS is not on this record yet. There is no CISA KEV listing. No Exploit-DB or Metasploit reference is linked. Use the source links for the original advisory.

Technology hubs

Directus security

Code

The project this record applies to. Exploit listings stay in References.

Timeline

Detection, updates, and resolution for this record.

  1. Oct 19, 2026 · Maintainer security tabs

    Detected

    ### Summary The AI Chat endpoint ( ) did not enforce the admin-only restriction on AI tools. Any authenticated user with app access could request the , , , and tools, which are reserved for administrators. The MCP endpoint enforced this restriction correctly. ### Impact The tools ran with the requesting user's own accountability, so they did not grant permissions beyond what the user already had:

Sources

  • Maintainer security tabsVendor advisoryGHSA-xf86-hfpm-92xr

    ### Summary The AI Chat endpoint ( ) did not enforce the admin-only restriction on AI tools. Any authenticated user with app access could request the , , , and tools, which are reserved for administrators. The MCP endpoint enforced this restriction correctly. ### Impact The tools ran with the requesting user's own accountability, so they did not grant permissions beyond what the user already had:

    github.com

Freshness

Last verified in this catalog Oct 19, 2026. Significant claims (KEV, EPSS, CVSS, affected ranges) come from the source rows above — open the original publisher for the authoritative advisory.

  • Maintainer security tabsNov 1, 2026