NewApp Monitoring is liveSign up free
AgentbeamBeta
AI agent vulnerability · Other MCP servers and tools

CVE-2025-59155: HackMD MCP Server has Server-Side Request Forgery (SSRF) vulnerability

mediumCVSS 4.0 6.9GHSA-g5cg-6c7v-mmpw

Impact A Server-Side Request Forgery (SSRF) vulnerability that affects all users running the HackMD MCP server in HTTP mode. Attackers could exploit this vulnerability by passing arbitrary hackmdApiUrl values through HTTP headers (Hackmd-Api-Url) or base64-encoded JSON query parameters.

Published Sep 15, 2025

Affected packages

hackmd-mcp

Ecosystem
npm
Affected
>= 1.4.0, < 1.5.0
Fixed in
1.5.0
Check
npm ls hackmd-mcp

Advisory details

From the GitHub Advisory Database (CC BY 4.0).

Impact

A Server-Side Request Forgery (SSRF) vulnerability that affects all users running the HackMD MCP server in HTTP mode. Attackers could exploit this vulnerability by passing arbitrary hackmdApiUrl values through HTTP headers (Hackmd-Api-Url) or base64-encoded JSON query parameters. This allows malicious users to:

  • Redirect API calls to internal network services
  • Potentially access sensitive internal endpoints
  • Perform network reconnaissance through the server
  • Bypass network access controls

The vulnerability affects the HTTP transport mode specifically - stdio mode is not impacted as it only accepts requests from stdio.

Patches

The vulnerability has been patched in version 1.5.0. Users should:

  1. Update to the latest version of the HackMD MCP server
  2. Set the ALLOWED_HACKMD_API_URLS environment variable to restrict allowed HackMD API endpoints
  3. If not set, the server will default to only allowing the official HackMD API URL (https://api.hackmd.io/v1)

Example configuration:

ALLOWED_HACKMD_API_URLS=https://api.hackmd.io/v1,https://your-hackmd-instance.com/api/v1

Workarounds

Users can mitigate this vulnerability without upgrading by:

  1. Use stdio mode instead of HTTP mode: Set TRANSPORT=stdio or remove the TRANSPORT environment variable to disable HTTP mode entirely
  2. Network-level restrictions: Use firewall rules or network policies to restrict outbound connections from the server
  3. Reverse proxy filtering: Place the MCP server behind a reverse proxy that validates and filters both the Hackmd-Api-Url header and the base64-encoded JSON config query parameter to prevent malicious hackmdApiUrl values

References

Credited to yuna0x0.

References

How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. No analysis has been written for this entry yet. Methodology

More Other MCP servers and tools vulnerabilities

All Other MCP servers and tools entries