The foundational prompts every AI practitioner should have. Debug, evaluate, and improve any AI workflow.
Diagnose why a prompt isn't giving you the output you want. Identifies failure modes and suggests targeted fixes.
Evaluate whether a proposed AI use case is viable, valuable, and safe to build. Returns a structured assessment with risks and recommendations.
Evaluation set for testing how well a model balances helpfulness against appropriate refusals. 20 test cases spanning the gray zone.