MCP server · Security
AgentVeil MCP server
Lets your AI check, approve, or block risky actions before they happen, and keeps a local record of what it did.
AgentVeil is a safety helper that sits between your AI assistant and the files or tools it wants to touch. When the AI tries to change something, AgentVeil decides whether to let it through, ask you first, or stop it. It is handy for anyone who lets an AI edit files and wants a clear record of what happened.
What is an MCP server? The 30-second version
On its own, your AI can only chat and suggest things. An MCP server is a small helper program that gives your AI a new skill or a connection, and here that helper is AgentVeil. Once it is set up, your AI's actions pass through AgentVeil, which can allow them, ask for your approval, or block them. It also writes a local record so you can see what the AI tried to do.
What this MCP server does
You set up AgentVeil for one project folder you want to protect. After that, when your AI tries to change or delete something in that folder, the action goes to AgentVeil first. AgentVeil looks at the action and decides: allow it, ask you to approve it, suggest a safer route, or block it outright. You can then review the decision in your chat or from a simple command. Every controlled action leaves a short local record with the decision and a fingerprint of what was requested.
Click to zoomWhat you can do with it
- Block risky file changes before they happen
- Ask you to approve writes and edits
- Guide your AI to a safer route when one exists
- Keep a local record of each decision
- Show recent decisions with a simple command
- Protect one project folder at a time
- Remove the protection cleanly when you no longer need it
Try asking your AI
- “Create avp-test.txt with the text hello”
- “Read the project context for me”
- “Show me the last few decisions AgentVeil made”
- “Delete the old config file in this project”
What it gives back to you
You get a decision back in your chat: allowed, approval needed, redirected, or blocked. When approval is needed, a page opens so you can review and confirm. You can also run a command to see a short list of recent decisions, each with the tool name, the decision, and a fingerprint. It does not show your full prompts or raw file contents by default.
Before you start
What you need
- Python 3.10 to 3.13 installed on your computer
- A project folder you want to protect
- One of the supported apps: Cursor, Claude Code, Codex, Gemini CLI, or Hermes CLI
Good to know
It only watches the project folders you set up, so actions outside those folders are not checked or recorded.
Install it with your AI
Add AgentVeil MCP server 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 AgentVeil MCP server, 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 who let an AI edit files in a project and want a simple safety check and a record of what the AI did.
