Skip to main content

About

TrustedMCP is a directory of MCP (Model Context Protocol) servers: the connectors that let AI assistants such as Claude, Cursor and VS Code work with tools like GitHub, Slack, Stripe or a database. It's run by a small team that also builds MCP servers for clients, so we have a stake in this topic.

How the grades work

Instead of ranking servers by popularity, each listing gets an editorial A-D grade built only from signals anyone can look up:

  • Login: does the server support or require authentication?
  • Upkeep: how recently and how often is it updated?
  • Known bugs: are there published vulnerabilities, and are they fixed?
  • Connection: does it run locally or over HTTP, and what are the defaults?

The grade is an assessment of those public signals, not a security audit or a certification. A good grade doesn't mean a server is safe in your setup. Use it as a starting point for your own review.

Grades are re-checked over time, and each page shows when it was last updated. Listings also show community reviews from people who have run the server, kept separate from the grade.

What's on trustedmcp.io

  • MCP server directory: each server with its grade and why it got it.
  • State of MCP server security: our report on the official MCP Registry, with the method, the limits and the raw data.
  • Registry checks: the official MCP Registry checked each night against public vulnerability databases (OSV, GitHub advisories, NVD).
  • Guides on MCP security, authentication, prompt injection and setting up popular servers.
  • Check your MCP stack: pick the servers you run and see each one's grade and weakest check. No signup, nothing stored.

About this blog

This is where we post our research and longer write-ups. When a post first appeared on trustedmcp.io or elsewhere, it links to the original.

Corrections

If we got a grade or a number wrong, tell us through the contact page and we'll fix it.