{
  "context": "Evaluates how well the solution leverages the punycode package to turn a single Unicode domain label into its ASCII Punycode form while preserving ASCII-only input and surfacing dependency errors. Scoring prioritizes direct use of the library API and correctness of the encoded outputs called out in the spec.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "punycode usage",
      "description": "Imports and calls `punycode.encode` on the provided label instead of reimplementing the encoding algorithm or using unrelated APIs, ensuring the dependency performs the conversion.",
      "max_score": 35
    },
    {
      "name": "mañana result",
      "description": "Encoding the label \"mañana\" returns exactly \"maana-pta\" as produced by `punycode.encode`.",
      "max_score": 20
    },
    {
      "name": "bücher result",
      "description": "Encoding the label \"bücher\" returns exactly \"bcher-kva\" as produced by `punycode.encode`.",
      "max_score": 20
    },
    {
      "name": "ASCII passthrough",
      "description": "An ASCII-only label such as \"example\" is returned unchanged, matching `punycode.encode` behavior without extra modifications.",
      "max_score": 15
    },
    {
      "name": "RangeError propagation",
      "description": "Any RangeError raised by `punycode.encode` for invalid or overflowing input is allowed to surface (or is rethrown unchanged) instead of being swallowed or replaced with a different error type.",
      "max_score": 10
    }
  ]
}