andrefetch/postal ? reverse-engineered prompt

Reverse engineered prompt

Build me a terminal based AI coding assistant called Postal that can work inside my repo and help me plan, read files, edit code, run shell commands, and review changes.

I want it to feel like a full screen chat app in the terminal, with live streaming responses, visible tool actions, and a clean status area for things like token usage and current model. It should let me log in with OpenRouter, switch models during a session, and run in both interactive mode and a simple one shot mode from the command line.

Please make it safe by default, so any risky or mutating action asks for approval based on a policy I can change. I also want saved sessions, continue and resume support, checkpoints, and a rewind command so I can back up if something goes wrong.

If you need to look up current docs online while building it, go ahead.

Are you gonna build this?

make sure you review the code using coderabbit

Try freeSponsored — opens CodeRabbit in a new tab