Azure-Samples/foundry-citadel-platform — reverse-engineered prompt

Build me an Azure based starter platform for governing and operating AI agents at enterprise scale. I want something like a Foundry Citadel reference setup that gives teams a fast path to launch agents, but with strong guardrails from day one.

The experience should feel like one click deployment for a shared governance hub plus separate team environments. It should enforce runtime policies like identity checks, rate limits, content safety, and cost tracking. I also want a control plane view that shows tracing, observability, compliance status, evaluations, and fleet level dashboards across all agents. Add an agent identity model so every agent has ownership, lifecycle controls, and access rules, including a way to handle shadow AI. Include security and data protection across the whole flow, plus a shared registry for approved agents and patterns.

Please make it production minded, clear, and adaptable rather than overly custom. Good docs matter here, so explain the architecture, setup, and extension points in plain English. Look up current Azure docs if you need to.