FrancescoStabile/numasec — reverse-engineered prompt

Build me an open source terminal app called numasec that acts like an AI security assistant for authorized security work. I want to run it from a project folder, choose a mode like AppSec, pentest, OSINT, lab, or research, and have it help me follow security runbooks using the tools already installed on my machine.

It should keep the whole operation organized, including target, scope, notes, commands, findings, evidence, replay steps, and report status. Findings shouldn’t just be chat messages, they should be saved with severity, proof, next action, and whether they are reportable. I should be able to resume old operations, rename them, export a share bundle, and generate a useful report at the end.

Make it feel like a smart terminal workspace, not just a scanner wrapper or chatbot. Use TypeScript for the CLI, keep the setup simple with npm install and numasec to start, and look up current docs online if needed.