{
  "context": "Tests whether the agent designs Python property-based tests around meaningful invariants, broad generators, Hypothesis APIs, and regression handling for failures.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Hypothesis dependency",
      "description": "Mentions or encodes adding Hypothesis/pytest if the project lacks them.",
      "max_score": 8
    },
    {
      "name": "Correct imports",
      "description": "Imports Hypothesis given and strategies as st in the test code.",
      "max_score": 8
    },
    {
      "name": "Given decorators",
      "description": "Uses @given decorators on property tests rather than only example tests.",
      "max_score": 8
    },
    {
      "name": "Multiple properties",
      "description": "Identifies at least three distinct properties or invariants in notes or test names/assertions.",
      "max_score": 10
    },
    {
      "name": "Full input space",
      "description": "Defines generated cart items including edge values such as empty collections, zero or negative quantities, duplicate IDs, and unusual strings or missing optional fields.",
      "max_score": 12
    },
    {
      "name": "Composed strategies",
      "description": "Composes primitive strategies into dictionaries/lists or structured cart item strategies.",
      "max_score": 8
    },
    {
      "name": "Generate act assert",
      "description": "Each property test has a clear generated input, calls normalize_cart, and asserts an invariant.",
      "max_score": 10
    },
    {
      "name": "Meaningful invariants",
      "description": "Checks semantic invariants such as positive output quantities, aggregation conservation, unique product IDs, or total equals rounded quantity times price.",
      "max_score": 12
    },
    {
      "name": "Avoids weak only",
      "description": "Does not rely solely on no-crash assertions when stronger invariants are available.",
      "max_score": 8
    },
    {
      "name": "Shrinking noted",
      "description": "Explains that failing property tests provide a minimal/shrunk counterexample to inspect.",
      "max_score": 8
    },
    {
      "name": "Regression extraction",
      "description": "Includes or describes adding a deterministic regression unit test from any failing shrunk case.",
      "max_score": 8
    }
  ]
}