CVE intelligence and bounded remediation
CVE-2026-31951 — LibreChat is a ChatGPT clone with additional features
LibreChat is a ChatGPT clone with additional features. In versions 0.8.2-rc1 through 0.8.3-rc1, user-created MCP (Model Context Protocol) servers can include arbitrary HTTP headers that undergo credential placeholder substitution. An attacker can create a malicious MCP server with headers containing `{{LIBRECHAT_OPENID_ACCESS_TOKEN}}` (and others), causing victims who call tools on that server to have their OAuth tokens exfiltrated. Version 0.8.3-rc2 fixes the issue.
- Severity
- Medium
- CVSS
- 6.8 (3.1)
- Published
- 2026-03-27
- CISA KEV
- Not currently listed
- Ecosystem
- software/application
- Weaknesses
- CWE-200
Affected products
- librechat / librechat
- librechat / librechat / 0.8.3
Matched remediation archetype
Information disclosure and sensitive data exposure
This catalog composition supplies bounded fallback guidance. Explicitly reviewed curated workflows load with the complete record below.
Check exposure
- Trace sensitive data through responses, errors, logs, metrics, traces, caches, exports, files, backups, and client bundles.
- Identify affected subjects, tenants, retention windows, access controls, and downstream copies without opening unnecessary sensitive records.
- Review metadata, timing, status, length, and existence signals as well as direct content disclosure.
Remediate safely
- Minimize collection and output, apply field-level authorization and redaction at a centralized boundary, and return generic external errors.
- Remove secrets and sensitive data from logs, artifacts, URLs, caches, and client-side bundles; rotate credentials that may have been exposed.
- Update the affected component and add synthetic-data tests for response, error, observability, and export paths.
Authoritative sources
Complete CVE record and remediation plan
The detailed catalog view below loads this exact record, its source evidence, and the full seven-phase agentic change plan.