H-mmer/pentest-agents — reverse-engineered prompt

Build me a Python bug bounty assistant suite for authorized testing. I want it to plug into Claude Code, Codex, Gemini, Cursor, Windsurf, VS Code Copilot, and OpenClaw, with ready made agents and commands that help a researcher run recon, find attack surface, hunt for issues, validate findings, connect related bugs into exploit chains, write reports, check duplicates, and learn from past work.

It should be able to scaffold a new bounty workspace for a platform and target, sync program details from places like HackerOne and Bugcrowd where possible, keep a persistent brain of endpoints and notes, and support autonomous hunt loops that still force validation before calling something real.

Please include MCP servers for bounty platform access and optional writeup search, a command line installer that can install, verify, render, and uninstall cleanly for each supported AI tool, plus tests so generated provider bundles don’t drift. Look up current docs online if you need to.