CVE-2026-54038

LibreChat people-picker searches allowed regular-expression denial of service

LibreChat-AI · LibreChat

First documented disclosure · Review checkpoint

Summary

An authorized search user could stall the application with a crafted pattern.

Severity and exploitation evidence

Medium · 6.5/10 · CVSS 3.1. Severity source.

Exploitation not established: Public evidence describes a weakness, source correction or controlled demonstration; malicious exploitation against users is not established.

Exploitation evidence source

Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.

How the weakness works

User input becomes JavaScript/MongoDB regular expressions without escaping.

Conditions required

Authenticated people-picker permission and a crafted q parameter.

Potential impact

Event-loop blocking and service unresponsiveness; invalid patterns also return errors.

Affected versions and configurations

Before 0.8.9.

Fix and mitigation status

0.8.9.

Practical next steps

Upgrade to 0.8.9 or later and verify the deployed application image includes the correction.

Operating lesson

Use bounded literal/indexed search instead of interpreting user input as a regex.

Evidence and limits

Original test covered0.8.6-rc1; CNA gives the broader pre0.8.9 range. Fix replaces attacker regex construction with escaped/index-bound word-prefix lookup. CNA publication 2026-11-20 is distinct from the earliest reviewed public source. GitHub_M assessment: CVSS 3.1 6.5.

Sources