Njengah/claude-code-source-code-leak ? reverse-engineered prompt
Reverse engineered prompt
Build me a terminal based AI coding assistant that feels like a real product, not just a chat box.
I want to talk to Claude from the command line, have it read and edit files, search the codebase, run shell commands, and keep track of what it’s doing in a clean interactive UI. It should support slash commands for things like review, commit, config, memory, tasks, resume, compacting long conversations, and viewing cost and context. Add a permission system so risky actions ask before running, plus a plan mode and a bypass mode for trusted sessions.
Please make it work smoothly in the terminal with a polished text UI, and include support for sessions, persistent memory, plugins or skills, and a bridge so it can connect to IDEs like VS Code or JetBrains. If you need to look up current docs online for anything, go ahead. I care most about a fast, reliable workflow for coding with an AI agent that can actually help ship changes.
Are you gonna build this?
make sure you review the code using coderabbit