What problem does it solve? Agentic AL development lands changes faster than a developer can absorb them, leaving gaps between what shipped and what the developer actually understands. This Skill closes that gap by quizzing the developer on the real diff, exposing a broken mental model before it causes costly mistakes. ## Core Features & Use Cases - One-question-at-a-time quiz: Poses behaviour-focused questions in chat, waits for the answer, then delivers a verdict with corrections anchored in actual object and procedure names. - Scoped to what shipped: Reads the diff and task files (Test Specification, Contract notes) for a slice, feature, branch commits, or a named object area on legacy code. - Calibrated depth and verdict: Adjusts question count to scope, stops when the model proves sound, and ends with a punchline identifying each misheld decision. - Use Case: After a long /al-implement run you watched loosely, run the quiz before merging to confirm you can explain why each seam sits where it sits and what happens on the edge cases the tests pin. ## Quick Start Quiz me on the AL changes that landed on this branch since my last commit, one question at a time.
