MCP server · Security
MCP Audit proxy
by P4ST4S
Adds a safety layer between your AI and its tools, logging every action and blocking risky ones.
MCP Audit is a small program that sits between your AI assistant and the tools it uses. It watches every request, writes it down in a log, and can block or slow down things you do not want. It is handy if you use AI at work and someone needs to know what the AI actually did.
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, like reading files or calling another app. This one is a bit different: it is a helper that wraps around another helper. You put it in the middle between your AI and, say, a file-reading tool, so every action passes through it and gets recorded.
What this MCP server does
You ask your AI to do something, like read a file. Your AI sends that request to MCP Audit instead of straight to the tool. MCP Audit writes the request into a log file, checks it against your rules, and then passes it on to the real tool. When the answer comes back, MCP Audit logs that too, hides sensitive bits if you asked it to, and hands the result to your AI. You see the normal answer in your chat, and a record of what happened sits in the log.
Click to zoomWhat you can do with it
- Keep a signed log of every tool call your AI makes
- Block tools you do not want the AI to use
- Limit how often a tool can be called per minute
- Hide sensitive fields like passwords before they are stored
- View recent activity in a small local dashboard
- Export activity as metrics for monitoring tools
- Wrap any existing MCP server without changing it
Try asking your AI
- “Read the file notes.txt in my temp folder”
- “List the files in my project directory”
- “Show me the last few audit entries from today”
- “Summarize which tools my AI called this morning”
What it gives back to you
Your AI still answers you normally, as if nothing changed. Behind the scenes, MCP Audit writes each request and response into a log file or a small database. You can open the local dashboard in your browser to see recent entries, top tools, calls today, and error rate. The log entries can be signed so you can tell if someone tampered with them.
Before you start
What you need
- The MCP Audit program itself (a prebuilt file you download, or Docker)
- A secret code (a long random string) used to sign the log
- An MCP server you want to watch, for example a file-reading tool
- A config file if you want to change the default settings
Good to know
It sits in the middle of your AI traffic and can block or slow down tool calls, and it stores a copy of requests and responses in a log, so make sure the log file is kept somewhere safe.
Install it with your AI
Add MCP Audit proxy 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 Audit proxy, 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.
Who it's for
People in regulated or careful workplaces, like IT, security, compliance, and operations teams, who need to know what their AI assistant actually did.
