smtg-ai/claude-squad — reverse-engineered prompt

Reverse engineered prompt

Build me a terminal app called Claude Squad that lets me run and manage multiple AI coding agents at the same time from one screen. I want to be able to spin up separate sessions for things like Claude Code, Codex, Gemini, or Aider, give each one its own task, and let them work in the background while I keep an eye on everything in a simple terminal UI.

Each task should use its own isolated git workspace and branch so the agents do not step on each other. I want to be able to open a session, send follow up prompts, pause and resume it, delete it, review diffs, checkout changes, and commit and push when I am happy. It would also be great to support an auto accept mode for agents that allow it.

Make it feel like a real tool I can install and run as cs, with sensible config for choosing which assistant to launch by default or via named profiles. If you need details, check the current docs online.

Want more depth? Deep Reverse