elder-plinius/T3MP3ST — reverse-engineered prompt

Build me a self hosted offensive security app called T3MP3ST that turns an AI coding agent into an authorized red team helper. I want a browser based War Room and a CLI where I can point it at a target I own or have written permission to test, then have it handle recon, exploitation attempts, and reporting in one flow. It should work with the agent I am already signed into, like Claude Code or Codex, without forcing new API keys, but also let me use API keys if I want, and support fully local models too.

Please keep the current scope honest. Web app testing and CTF style workflows should feel solid first. Source code analysis can be limited, and cloud, mobile, and binary features can exist as early scaffolding instead of pretending they are fully mature. I also want reproducible benchmark and claim checking built in, plus clear docs, examples, and a simple setup so I can run the server locally and start from the UI fast. Look up current docs online if you need to.