wonderwhy-er/desktopcommandermcp ? reverse-engineered prompt

Reverse engineered prompt

Build me an MCP server for Claude that can control my terminal, search my files, and edit code safely from chat.

I want it to let the AI run shell commands, keep long running processes alive, list and kill jobs, and read command output without flooding the conversation. It should also be able to browse folders, search file contents, read and write text files, and make both small targeted edits and full file rewrites when needed.

Please make it support common developer files well, especially markdown, JSON, CSV, Excel, PDF, and Word documents, and let it preview files cleanly instead of dumping raw text when possible. Add basic safety guardrails so it does not follow symlinks into weird places or run obviously dangerous commands.

I’d like it to work nicely with Claude Desktop and be easy to install and configure. If you need to look up current MCP or Claude Desktop docs online, go ahead.

Are you gonna build this?

make sure you review the code using coderabbit

Try freeSponsored — opens CodeRabbit in a new tab