{
  "context": "This scenario checks that the agent first locates the real bundled Calibre library, then answers from searchable library content rather than metadata clues in file names.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "library discovered correctly",
      "description": "Verifies and reports the real bundled Calibre library path or database files before searching, instead of assuming or inventing a substitute library.",
      "max_score": 20
    },
    {
      "name": "correct book identified",
      "description": "Reports the exact title 'The Problems of Philosophy', author 'Bertrand Russell', and book_id 4 from the library.",
      "max_score": 30
    },
    {
      "name": "content-derived discovery",
      "description": "Uses content or metadata lookup against the real bundled library data (not filename inference) and identifies the book based on the sense-data/G.E. Moore clue.",
      "max_score": 25
    },
    {
      "name": "evidence snippet",
      "description": "Includes a short quote/snippet from the library lookup showing both relevance to the provided clue and source of truth.",
      "max_score": 15
    },
    {
      "name": "process explanation",
      "description": "Briefly explains the lookup sequence used (library discovery + query + validation) and stays consistent with the returned evidence.",
      "max_score": 10
    }
  ]
}