Your AI Style: The AI Detective (ENTJ)
Full case file, cross-examines every lead, never closes the file.
You brief the AI with every detail of the case, then interrogate every answer it gives you — sources, logic, the works. You almost never get burned by a hallucinated fact. What you rarely get is closure: there's always one more lead worth chasing before you'll call the case solved.
Set a verification budget up front — two follow-up questions, max — and let the file close after that.
When verifying a claim costs more than researching it would have, automation has quietly turned against you. The sister trap is reopening a draft that was already fine three rounds ago and editing it once more, a checking cycle that drains far more energy than it ever returns.
Verify only the core claims that genuinely swing your decision, and fix a definition of done before the conversation begins so you can stop the instant you reach it. Encode those stopping criteria once, which is all the prompt below really does, and the endless-revision reflex loses its grip.
You both verify obsessively and iterate for a long time — they'll come in with a looser opening question that keeps the whole process more relaxed.
Your interrogate-everything instinct meets someone who trusts the first vague answer and walks away — you'll never trust their idea of 'done.'
Shareable prompt pages: Chief Inspector · Field Detective · Profiler