manashbajpai8840-alt/ai-soc-simulator ? reverse-engineered prompt

Build me a safe AI powered SOC simulator that generates synthetic security events and shows them in a live dashboard.

I want it to create mixed traffic like brute force SSH attempts, SQL injection, malware style C2 beacons, and normal benign logs, then run a simple first pass to catch obvious suspicious events before sending only the flagged ones to Claude for triage. If there is no API key, it should still work in a mock mode and make believable threat decisions so I can demo it without setup.

When the system thinks something is a real threat, it should simulate a response like blocking an IP, isolating a process, and writing a forensic report, while staying safe by default so nothing changes on my machine unless I explicitly turn on live mode. Also store the events, verdicts, blocks, and reports in SQLite, and show a Streamlit dashboard with live counts, threat breakdowns, active blocks, and incident details.

If you need current docs or best practices for any library, look them up online.