{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/5ae859759c7e4036a5c73b251164e3e9\" frameborder=\"0\" width=\"1440\" height=\"1080\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1080,"width":1440,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1080,"thumbnail_width":1440,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/5ae859759c7e4036a5c73b251164e3e9-c35495103fc1443d.gif","duration":565.804,"title":"SOC Triage AI Project Overview and Results 🚀","description":"Hey, I am Solomon Smith, and I built a SOC triage AI. It loads 11 docs at startup, pulls the four most relevant pieces for each alert, and if nothing matches well enough it stops and tells the analyst to check manually. When it finds a match, it sends the alert plus the evidence to Claude, which returns a fixed format report with severity, escalation, MITRE techniques, and reasoning. In the demos it correctly identified LockBit ransomware and credential theft, and the guardrail test used gibberish to ensure it did not guess. I tested seven scenarios, hit 7 out of 7 passing, with about a seven second response time, and I did not request viewer action."}