MCP server · Security · official
DashClaw MCP server
by ucsandman · official
Your AI has to ask you before it runs risky commands, and every decision is written down.
DashClaw is a safety layer that sits between your AI assistant and the things it wants to do. When the AI is about to run something risky, like deleting files or force-pushing code, DashClaw stops it and asks you first, even if you are away from your desk. It is handy for anyone who lets an AI work on its own for long stretches and does not want to babysit every single step.
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 another service. This one connects your AI to DashClaw, a safety checkpoint that checks each action against your rules before it happens. So when you ask your AI to do something, DashClaw can pause it, ask for your yes or no, and keep a record of what was decided.
What this MCP server does
You set up DashClaw and tell it what your rules are. Then you connect your AI assistant to it. When the AI wants to run a command, DashClaw looks at that command first. If it looks dangerous or expensive, DashClaw freezes it and sends you a request for approval, which you can answer from your phone, email, or chat app. Once you approve or deny, DashClaw writes down the decision so you can look back later and see exactly what happened.
Click to zoomWhat you can do with it
- Stop a risky command before it runs and ask you for approval
- Approve or deny requests from your phone or chat app
- See a plain-English explanation of what each pending command would do
- Set rules that say which actions always need your approval
- Review a log of every decision that was made
- Allow certain safe actions to run without asking you each time
- Check whether the safety hooks are working properly
Try asking your AI
- “I am going to run a command that deletes the build folder. Can you check if that is allowed?”
- “Please ask DashClaw if I can force-push to the main branch before you do it.”
- “Show me the last ten actions that needed approval and what I decided.”
- “Set up a rule that any command touching the production database needs my approval.”
What it gives back to you
When you ask your AI to do something, DashClaw either lets it through, warns you, or stops it and asks you. If it needs your approval, you get a message with a short explanation of what the command does and a button to approve or deny. After you decide, the chat shows the outcome and a note that the decision was recorded. You can also ask for a summary of recent decisions and see a list of what was allowed, blocked, or approved.
Before you start
What you need
- A computer or server where you can run DashClaw
- Node.js installed (the free tool that runs JavaScript programs)
- An AI assistant that supports hooks or MCP, like Claude Code or Codex
- A way to receive approval requests, such as a phone or chat app
Good to know
DashClaw can stop your AI from doing things, but it cannot protect against an AI that has permission to change DashClaw itself, so keep it in a separate account or container if you need stronger safety.
Install it with your AI
Add DashClaw 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 DashClaw 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
Solo developers and small teams who run AI coding agents unattended for hours and want a safety net.
