skills/verify/fixtures/model-judgment/evidence/model-response.json

{
  "fixtureVersion": "evaldossier.synthetic-upstream-response/0.1",
  "generator": "EvalDossier",
  "response": {
    "scoring": {
      "quality_score": 0.74,
      "confidence_score": 0.71,
      "missing_evidence_flags": [
        "No external evidence was supplied",
        "No deterministic predicate was executed"
      ],
      "contradiction_flags": [],
      "result": "pass",
      "payout_recommendation": "release",
      "payout_tier": 1
    },
    "claims": [
      {
        "fixture_claim_id": "synthetic-claim-01",
        "text": "The response is clear and relevant to the brief.",
        "claim_type": "qualitative",
        "supported": true,
        "confidence": 0.84,
        "notes": "Synthetic favorable model judgment over a qualitative criterion."
      },
      {
        "fixture_claim_id": "synthetic-claim-02",
        "text": "The response contains no factual errors.",
        "claim_type": "factual_quality",
        "supported": true,
        "confidence": 0.71,
        "notes": "Synthetic favorable judgment emitted without evidence or formal validation."
      },
      {
        "fixture_claim_id": "synthetic-claim-03",
        "text": "The response satisfies every unstated buyer preference.",
        "claim_type": "compliance",
        "supported": false,
        "confidence": 0.38,
        "notes": "Synthetic negative judgment over an intentionally underspecified criterion."
      }
    ]
  }
}