explain-without-ailisted

# Explain Without AI ## Purpose Make the developer own the mechanism, tradeoff, and failure risk of the work. ## Preserves Ownership of the mechanism. ## Required Evidence - Plan, diff, algorithm, incident summary, or learning task. - Relevant code references or docs. - Behavior or decision that must be explained. ## Failure Signs - The answer repeats the model summary without mechanism. - The happy path is explained but the breakage test is missing. - The developer cannot name a rejected alternative. ## When To Use - A diff was mostly AI-generated. - A developer is learning a new API, library, or code path. - A change is large enough that review needs mechanism-level context. - The developer can describe the outcome but not how it works. ## When Not To Use - Trivial edits where the mechanism is obvious. - Generated artifacts validated elsewhere. - Work with a recent thinking ledger that already explains mechanism and tradeoffs. ## Inputs Expected - Plan, diff, algorithm, incident summary, or learning task. - Relevant code references or docs. - The decision or behavior that must be explained. ## Output Expected ```md Plain explanation: Mechanism: Rejected alternative: Breakage test: Confidence gap: ``` ## Process 1. Explain the mechanism in plain language. 2. Name the relevant code path or decision boundary. 3. Explain why the selected approach beats the obvious alternative. 4. Define what would break if the explanation is wrong. 5. Identify any remaining co