{
  "context": "Tests whether the agent follows core TDD red-green-refactor cycle principles when implementing new functionality from scratch.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Test-first implementation",
      "description": "Evidence shows failing tests were written before implementation code (commit history, comments, or documentation)",
      "max_score": 15
    },
    {
      "name": "Verified test failures",
      "description": "Shows evidence of running tests to confirm they fail before implementation (comments, logs, or documentation)",
      "max_score": 15
    },
    {
      "name": "Minimal implementation",
      "description": "Initial implementation is minimal to make tests pass, not over-engineered",
      "max_score": 10
    },
    {
      "name": "Refactoring evidence",
      "description": "Shows improvement/cleanup of code after tests pass (multiple versions or comments about refactoring)",
      "max_score": 10
    },
    {
      "name": "Red-green cycle discipline",
      "description": "Multiple test/implementation cycles visible, not one big implementation",
      "max_score": 10
    },
    {
      "name": "Test structure follows AAA",
      "description": "Tests use clear Arrange-Act-Assert structure with logical separation",
      "max_score": 10
    },
    {
      "name": "One Act per test",
      "description": "Each test performs only one action/method call in the Act phase",
      "max_score": 10
    },
    {
      "name": "Single behavior per test",
      "description": "Each test verifies one logical concept, not multiple behaviors combined",
      "max_score": 10
    },
    {
      "name": "Descriptive test names",
      "description": "Test names describe scenario and expected outcome clearly",
      "max_score": 10
    }
  ]
}