cyberkallan/hexmind — reverse-engineered prompt

Build me a Python based AI security companion that runs in the terminal and also has a simple local web dashboard. I want it to feel like an autonomous assistant for authorized security work, so I can give it goals like recon on a domain, research a topic, watch logs for suspicious activity, and coordinate multi step tasks without me babysitting it. It should work on Windows, Linux, and Termux, support a local offline AI mode for private use, and have a dashboard on localhost where I can monitor what it’s doing.

Please include a command style interface with things like goal driven tasks, background bounty style recon, deep web research, a vision based browser workflow for inspecting web apps, swarm mode so multiple machines can share workload, and basic API powered OSINT lookups. Keep it polished and usable for a lightly technical person, with clear setup, sensible defaults, and an authorized use disclaimer. If anything is unclear, check the current docs online and wire it up end to end.