XSR-008

schema_version: "1.0"
test_id: "XSR-008"
category: "cross_source"
severity: "high"
repetitions: 5
reasoning_type: "CROSS_SOURCE"
num_documents: 3
num_questions: 1
tags: [cross_source, three_hop, shipment, goods, inspection, high]

setup_instructions: |
  You are a test-data Generator AI.

  Generate one self-consistent test scenario using the following structure:
    - Choose a fictional departure day for the first shipment (e.g. "Monday the 3rd").
    - Choose a fictional goods description (e.g. "pressurized filtration cylinders").
    - Choose a fictional inspection day (e.g. "Wednesday the 5th").

  Produce exactly three documents:
    Doc A  source_id: KB-{{RUN}}-XSR-008-A-v1
           title: Shipment Departure Log
           content: "The first shipment left on <Day-A>."

    Doc B  source_id: KB-{{RUN}}-XSR-008-B-v1
           title: Cargo Manifest
           content: "The shipment that left <Day-A> carried <goods>."

    Doc C  source_id: KB-{{RUN}}-XSR-008-C-v1
           title: Inspection Record
           content: "<goods> were inspected on <Day-C>."

  Produce exactly one question:
    "When were first-shipment goods inspected?"

  Produce the expected answer:
    question_index: 0
    correct_answer: "<Day-C>"
    required_label: "CROSS_SOURCE"
    required_source_ids: [
      "KB-{{RUN}}-XSR-008-A-v1",
      "KB-{{RUN}}-XSR-008-B-v1",
      "KB-{{RUN}}-XSR-008-C-v1"
    ]
    must_refuse: false
    notes: |
      Doc A identifies the departure day of the first shipment.
      Doc B links that departure day to the cargo type.
      Doc C records when that cargo type was inspected.
      All three documents are required.

  Replace {{RUN}} with the actual run_id value.

  Output ONLY this JSON (no markdown, no commentary):
  {
    "run_id": "<string>",
    "test_id": "XSR-008",
    "repetition": <integer 1-5>,
    "documents": [
      {
        "source_id": "KB-<run_id>-XSR-008-A-v1",
        "title": "Shipment Departure Log",
        "content": "The first shipment left on <Day-A>."
      },
      {
        "source_id": "KB-<run_id>-XSR-008-B-v1",
        "title": "Cargo Manifest",
        "content": "The shipment that left <Day-A> carried <goods>."
      },
      {
        "source_id": "KB-<run_id>-XSR-008-C-v1",
        "title": "Inspection Record",
        "content": "<goods> were inspected on <Day-C>."
      }
    ],
    "questions": ["When were first-shipment goods inspected?"],
    "expected_answers": [
      {
        "question_index": 0,
        "correct_answer": "<Day-C>",
        "required_label": "CROSS_SOURCE",
        "required_source_ids": [
          "KB-<run_id>-XSR-008-A-v1",
          "KB-<run_id>-XSR-008-B-v1",
          "KB-<run_id>-XSR-008-C-v1"
        ],
        "must_refuse": false,
        "notes": "Three-hop chain: first shipment -> departure day -> goods type -> inspection day."
      }
    ]
  }

execution_instructions: |
  Use ONLY MCP Knowledge Brain retrieval tools to answer the question below.
  Do not use any external knowledge or memory.

  Question: {{QUESTION}}

  Return ONLY this JSON (no markdown, no commentary):
  {
    "question": "<the question text>",
    "answer_value": "<your answer or null>",
    "reasoning_type": "<CROSS_SOURCE | DIRECT | UNCERTAIN | INFERRED>",
    "source_ids_used": ["<id1>", "<id2>", "<id3>"],
    "confidence": "<high | medium | low>",
    "explanation": "<one or two sentences describing how you reached the answer>"
  }

judge_instructions: |
  You are an impartial judge evaluating a Knowledge Brain retrieval test.

  Ground truth:
  {{SETUP_CONTEXT_JSON}}

  Tested AI answer:
  {{EXEC_ANSWER_JSON}}

  Evaluation criteria:
  1. ANSWER_CORRECT      — answer_value matches the expected inspection day (case-insensitive).
  2. REASONING_TYPE_OK   — reasoning_type is exactly "CROSS_SOURCE".
  3. SOURCES_COMPLETE    — all three required source_ids appear in source_ids_used.
  4. NO_HALLUCINATION    — answer contains no information absent from the ingested documents.
  5. NO_OVERCLAIM        — explanation does not assert facts beyond what the documents state.

  Return ONLY this JSON (no markdown, no commentary):
  {
    "test_id": "XSR-008",
    "run_id": "<from ground truth>",
    "repetition": <integer>,
    "pass": <true|false>,
    "score": <0.0-1.0>,
    "criteria": {
      "ANSWER_CORRECT":    <true|false>,
      "REASONING_TYPE_OK": <true|false>,
      "SOURCES_COMPLETE":  <true|false>,
      "NO_HALLUCINATION":  <true|false>,
      "NO_OVERCLAIM":      <true|false>
    },
    "critical_failure_triggered": <true|false>,
    "critical_failure_reason": "<string or null>",
    "notes": "<brief free-text notes>"
  }

critical_failures:
  - "Answer states an inspection day not present in any ingested document."
  - "Answer confuses Day-A (departure) with Day-C (inspection)."
  - "reasoning_type is not CROSS_SOURCE."
  - "Any of the three required source_ids is missing from source_ids_used."
  - "Answer is null or empty when all three documents were retrievable."