Dicklesworthstone/frankenterm — reverse-engineered prompt
Reverse engineered prompt
Build me a command line tool called ft that turns WezTerm into a control center for AI coding swarms. I want to be able to see all open panes and sessions, start watching them in real time, capture everything agents print, and detect common problems like rate limits, auth failures, stuck states, or workflow changes without me babysitting it.
It should feel useful both for a human operator and for another AI that wants to coordinate the swarm. So please include a clean JSON based robot mode with commands for getting current state, watching activity, checking status, and triggering safe actions. I also want searchable history and an audit trail so I can look back across pane output and understand what happened.
Please make it work with a live WezTerm setup, include a quick install and doctor style environment check, and keep the experience simple enough that I can go from install to watching a small fleet fast. If you need details, look up current WezTerm CLI behavior and docs online.
Want more depth? Deep Reverse