Advanced AI Query Library

Evidence-first diagnosis

Using only the evidence below:
A. list observed facts
B. list missing information
C. rank hypotheses with reasons
D. propose non-destructive tests
E. define escalation criteria
Do not invent missing telemetry.

Tutor mode

Teach me [topic] using progressive difficulty.
Start with a 5-question diagnostic.
Then create a lesson based on my gaps.
Use one worked example and one hands-on exercise.
Quiz me at the end without revealing answers first.

Compare options

Compare [A] and [B] for [specific use case].
Criteria: cost, learning curve, reliability, limitations, fit.
Separate facts from judgment.
State what information could change the recommendation.

Prompt debugging

If an answer is weak, diagnose the request: Was the goal ambiguous? Was key context missing? Were constraints contradictory? Was the requested output unverifiable? Improve the prompt rather than simply adding “be more detailed.”