TDD Guidelines

To enforce strict test-driven development, this skill requires agents to only write production code backed by a failing test that becomes passing evidence. It uses Red-Green-Refactor enforcement and compliance gates to prevent hallucinated or weak implementations.

The Red-Green-Refactor cycle requires writing a failing test first (Red), adding minimal code to pass it (Green), then cleaning up safely when tests are green (Refactor). This prevents code bloat by ensuring correctness and non-bloated implementations through repeatable workflows.

You log auditable FAIL to PASS test evidence by running tests through the TDD evidence logger script at the workspace root. This produces an evidence-logged validation history required before any disciplined commit checkpoints.

No, you cannot run tests in watch mode. The strict TDD agent workflow requires one-shot test execution to ensure reliable FAIL to PASS validation and generate accurate test evidence for compliance gates.

You isolate generated scripts and test outputs by sandboxing them under artifacts. This workspace isolation keeps temporary scripts out of the OSS template repo, ensuring tests don't pollute your codebase.

You need test-driven development evidence gates during coding, refactoring, and test-writing scenarios where correctness and repeatable workflows are required. They prevent weak implementations by producing an auditable FAIL to PASS history before commits.