I'm building a test rig for agent red team work, focusing on prompt injection against secured TEEs. I've got an Intel Xeon with TDX enabled in the BIOS and a fresh Ubuntu 22.04 install. The goal is to get a simple "hello world" agent running inside a TDX enclave and prove its integrity to a remote verifier.

I've seen the official Intel guides, but they're dense. For a practical attacker's perspective: what's the minimal path to get remote attestation working on my own hardware? I'm thinking:
- Which components are absolutely necessary (DCAP driver, PCCS, etc.)?
- Is there a way to simulate a local attestation service first, before setting up a full remote verifier?
- Common pitfalls in the setup that break the chain of trust?

My endgame is to understand the attestation flow well enough to later probe its weaknesses in a controlled environment.