Ten questions, five minutes, one number. The deployable agent stack has seven layers; this diagnostic scores the six that teams actually skip. It doesn’t test Layer 1 (the LLM, tools, and planning loop), because if you’re here you already have it, and it’s the layer everyone mistakes for the whole agent. An agent missing any of the other six is a demo. Find out which one you’ve built.
Score it against a real agent you run or plan to ship, not the one in the roadmap deck. Your score renders immediately; nothing is gated.
Two dropdowns that turn a pile of scores into something worth reading. Skip them and your result is unchanged.
Want the written report for your band, plus the five-article framework this diagnostic is built on? Subscribe to The Steel Thread — the score above stays yours either way.