Came across an interesting concept, instead of benchmarking the AI, it benchmarks *you*. Your agent answers questions about your working style, decision patterns, and how you treat it, then generates a personality archetype.

I ran it with my agent and got roasted pretty hard.

Has anyone done something similar? Would be interesting to compare results across different model backends.