momakkawi/agentic_tracking_system ? reverse-engineered prompt

Build me an AI driven attendance tracking system for a university or institution that turns noisy badge scan data into useful insights.

I want a web dashboard where admins can see attendance patterns, spot suspicious behavior like proxy attendance or duplicate scans, and explore student activity in a clear visual way. The backend should accept raw attendance data from common file formats, clean it up, remove duplicates, build sessions from scans, and combine that with class schedule data from calendar files. Then it should use a set of cooperating AI agents to validate the data, find groups or cohorts, generate insights, and answer questions in plain English.

Make it run with one command so it starts both the API and the dashboard together, and let the system reload config changes without restarting. Keep the interface modern and polished, with charts, timelines, and a strong focus on operational visibility. If you need current docs for any libraries, look them up online.