Identify and fix wrong ai behaviors

This skill is for Code Review & Testing and helps you set up evaluations for an llm application, create evaluation criteria for ai outputs, and build golden datasets for testing.