ShellWard MCP server

ShellWard is a safety helper for AI agents. It watches what your AI is about to do and steps in when something looks dangerous, like a command that could wipe your files or a message that leaks private information. It is handy if you use AI tools that can run commands, send emails, or touch files on your computer.

What this MCP server does

You ask your AI to run a command, read a file, or send a message. Before doing it, your AI asks ShellWard to check. ShellWard looks at the command, the text, or the file path and compares it against known bad patterns, like file-wiping commands or hidden instructions. It also scans for personal data such as phone numbers, ID numbers, and passwords. Then it tells your AI whether to go ahead or stop.

Click to zoom

What you can do with it

• Check if a shell command is safe before it runs
• Spot hidden prompt injection in text your AI reads
• Find personal data like phone numbers, ID numbers, and card numbers
• Check if a file path is sensitive, like .env or .ssh files
• Block tools that move money or transfer funds
• Scan an AI response for leaked secrets
• Run a compliance health check for Chinese rules like PIPL

Try asking your AI

• “Is it safe to run rm -rf on this folder?”
• “Check this customer file for personal information before I share it.”
• “Does this email text contain anything that looks like a prompt injection?”
• “Run a compliance check on this project.”

What it gives back to you

You get a clear yes or no answer, often with a short reason. For data scans, you get a list of what was found, like the type of personal data and where it appeared. For compliance checks, you get a red, yellow, or green scorecard. The answer shows up right in your chat, so you can decide what to do next.

Before you start

Who it's for

Anyone who uses AI agents that can run commands or handle files, especially people working with customer data or in regulated industries.