anpa1200/adversarygraph ? reverse-engineered prompt

Build me a self hosted cyber threat intelligence workbench that helps an analyst go from a report or IOC list to useful detections and hunting ideas.

I want to be able to paste in threat reports, PDFs, DOCX, or text, then have the app organize the evidence, map it to MITRE ATT&CK, enrich IOCs and CVEs, and show a clear reasoning trail from raw evidence to detection ideas. It should also include a threat radar view for early warning, a place to triage malware leads, a way to review asset exposure, and a simple attack simulation area for testing SIEM rules and validation results.

Make it run locally with Docker, keep it self hosted, and make the AI help draft suggestions without pretending it knows more than the evidence shows. If anything in the docs is unclear, look up the current docs online and keep the setup straightforward for someone who is not deeply technical.