All vulnerabilities

Servers and applications

GHSA-m73r-rpxg-gf5x

Possible regression of MCP OAuth callback browser-binding check

Severity
Medium
Published
17 Nov 2026
Last changed
17 Nov 2026
Added to UpHeron
17 Nov 2026
Fixed in
0.8.9
Weakness
Cross-site request forgery (CWE-352)
Sources
Project advisories (GitHub)

Description

Submitted privately as a possible regression of GHSA-vf7j-7mrx-hp7g. This is static source analysis only. No build, test, deployment, or runtime reproduction was performed. The report is for maintainer assessment of whether the later callback fallback reintroduces the previously disclosed condition, not a claim of a distinct root cause or runtime-confirmed impact.

The MCP OAuth callback accepts a callback without either of its browser-binding cookies when it can find a fresh PENDING MCP OAuth flow for the state value. The pending flow is created for the initiating LibreChat user. The callback then exchanges the provider authorization code and stores the resulting credentials using that flow's userId.

Consequently, an attacker who can initiate an MCP OAuth flow in their own LibreChat account may be able to send its authorization URL to a victim.

From LibreChat-AI/LibreChat's security advisory. Shortened; the full text is there.

Affected software

SoftwareAffected versions
LibreChatServers and applications · github:librechat-ai/librechat
  • Every version before 0.8.9

About LibreChat

Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active

From its GitHub repository.
Docker Hub pulls
669,094Docker Hub, 15 Nov 2026
GitHub stars
47,138GitHub, 15 Nov 2026
Forks
9,706GitHub, 15 Nov 2026
License
MITGitHub
First published on GitHub
12 Feb 2023GitHub

References