video1.0<iframe src="https://www.loom.com/embed/6402f69f6a774eeab29a83db67aabcd1" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/6402f69f6a774eeab29a83db67aabcd1-82af0002ba7dc16b.gif438.074This Loom presents AI-Pol, an AI Polygraph concept for an Autonomous Incident Response Agent for the Fine Evil Hackathon 2026. The system uses 13 models inspired by 30 animal defense and attack behaviors, with a Dreamcatcher-based theme for isolating “bad dreams” or cyber nightmares. In a Python simulation using Ollama with Llama, modules detect attacks such as missing authentication headers, brute force after three failed logins, and suspicious persistence, then generate an incident report with a score of 8 out of 10 and recommended SOC-TEM actions including isolation and logging to an immutable file. The example incident involves lateral movement and brute force, mapping tactics to T11-T12, privilege escalation T10-21 remote services, and exploit-related T11-90, with multiple isolations such as IP blocking and account lockdown.