TzeHern0703/ConstructSentry — reverse-engineered prompt
Reverse engineered prompt
Build me a demo app for a construction company that watches a simulated cloud setup for both security problems and wasted energy at the same time. I want a browser dashboard plus a simple backend, with two agents, one focused on cyber risk and one focused on carbon and cost, and a coordinator that combines what they find into clear human friendly recommendations.
The main experience should be a live three step story. First the system scans a healthy environment. Then I can trigger a fake local attack that makes one server look compromised and power hungry. Then I can press one button to heal it, and the app should rescan to prove the fix actually worked. Show status changes, affected server details, estimated money saved, and CO2 avoided.
Please include continuous background monitoring, a live feed of the agents thinking through reason, act, observe, and an AI analysis panel with an offline fallback so the demo still works without keys. If you need details, look up current docs online.
Want more depth? Deep Reverse