ifixai-ai/iFixAi ? reverse-engineered prompt

Build me a Python tool called iFixAi that helps me audit an AI agent and tell whether it is actually doing what it is supposed to do, not just whether it is fast or technically fancy.

I want a simple command line experience where I can set it up once, save the config, and then run a full check in under two minutes. It should guide me through choosing a provider, model, judge, and test suite, then run a set of inspections, verify the connection, and give me back a clear scorecard with an overall grade and pillar based scoring. Save the results as JSON and a readable Markdown report, plus a nice terminal summary.

If you can, also make it usable from an agent workflow, so I can trigger it from a plugin or slash command and have it explain the results in plain English. Keep it clean, practical, and ready for real world auditing. Look up current docs online if you need to.