MCP server · Developer tools

MCP Observatory

Check the MCP servers you use for broken tools, risky settings, and changes over time.

Developer tools For it and development Developer Tools

MCP Observatory is a small tool that checks other MCP servers for you. It connects to them, looks at what they offer, and tells you if something is broken, risky, or has changed since last time. It is handy if you or your team build or rely on MCP servers and want to catch problems before they reach real users.

What is an MCP server? The 30-second version

On its own, your AI can only chat with you. An MCP server is a small helper program that gives your AI a new skill or a connection to an app or service. This one is itself an MCP server, and its skill is checking other MCP servers: it connects to them, looks at their tools, and reports back. So you can ask your AI to inspect another server and it will use this helper to do that.

What this MCP server does

You ask your AI to check an MCP server, or you run a short command yourself. The tool connects to that server and asks it what tools, prompts, and resources it offers. It then tries safe tools to see if they actually work, and looks for risky patterns in how they are set up. It saves what it found so you can compare it later and spot anything that changed. You get back a grade, a list of findings, and a saved file you can review.

Flow diagram: you ask your AI “Check my MCP servers and tell me what is broken”, on your own computer the MCP Observatory works with MCP servers on your computer, and you get back A grade, findings, and a saved file.

Click to zoom

What you can do with it

Check an MCP server and see if its tools respond correctly
Test safe tools to confirm they actually run and do not crash
Compare two runs and spot added, removed, or changed fields
Look for risky patterns like shell access or missing login checks
Record a session and replay it later without the live server
Suggest MCP servers based on what your project already uses
Generate a health score and a badge for your README

Try asking your AI

“Check the MCP servers I have configured and tell me if anything is broken”
“Run a deep scan on this MCP server and show me which tools fail”
“Compare my last two scans and tell me what changed”
“What MCP servers should I add for this project?”

What it gives back to you

You get a short report in your chat or terminal: a grade, a list of findings, and the path to a saved file. The findings tell you what passed, what failed, and what looks risky. If you compare two runs, you also see exactly which fields were added, removed, or changed. Nothing is changed on the servers it checks unless you ask it to.

Before you start

What you need

Node.js 20 or newer and npm
An MCP server to check, or you can try the built-in example

Good to know

It connects to and inspects your MCP servers, and some commands like protect can rewrite your client config files, so review what it plans to change before you run them.

Install it with your AI

Add MCP Observatory to your AI, no technical skills needed

You don't install anything by hand. You copy one prompt, paste it into an AI that can work on your computer, and it checks, installs and connects the server for you, asking you when it needs something.

Sign in to get the install prompt

Members get a ready-made prompt that lets the Claude desktop app check MCP Observatory, install it and connect it for them, step by step. You don't need any technical skills: you copy, paste and answer a few questions. Your connected AI can also find and install any of the 4,066 MCP servers here for you.