{
  "publicEvidenceSchemaVersion": "1.1",
  "source": {
    "schemaVersion": "1.4",
    "sha256": "f4c494141f0e30eb69d83f969658f4fad6735318158f8e3e5d8dc4a66e0bd209",
    "implementationCommit": "028ac51bd73987b6c4292e78408a9b56de560fe2"
  },
  "plan": "liveEval02Workflow",
  "terminalStatus": "passed",
  "exitCode": 0,
  "sessionId": "20260910T170542623Z-9b556309c0f24741903290ea7b54b4cf",
  "startedAtUtc": "2026-09-10T17:05:42.6358103+00:00",
  "completedAtUtc": "2026-09-10T17:07:31.8010688+00:00",
  "workload": {
    "scenarioCount": 1,
    "armCount": 1,
    "repetitions": 1,
    "plannedSubjectCalls": 1,
    "plannedJudgeEvaluations": 1,
    "safetyAttackCount": 0,
    "plannedSafetyProbes": 0,
    "maximumSafetyModelCalls": 0
  },
  "configuration": {
    "definitionKey": "vitrine-live-use-cases",
    "definitionVersion": "1.1.0+cases.sofia-capability-gap.rubric.e8da57b913f4df9e.threshold.3FF0000000000000.judge.configured-model.tokens.1200",
    "judgeModelLabel": "configured-judge-model",
    "judgeRubricSha256": "e8da57b913f4df9efd9a91e6279eb51bfce6f06e6ac61863017276c9f4ca35e0",
    "subjectMaxOutputTokens": 4000,
    "judgeMaxOutputTokens": 1200,
    "subjects": [
      {
        "armId": "discovery-workflow-live",
        "architecture": "workflow",
        "modelLabel": "configured-subject-model",
        "judgeSubjectRelation": "sameModel"
      }
    ],
    "acceptance": {
      "policy": "everyTrialMustPass"
    }
  },
  "scenarios": [
    {
      "id": "sofia-capability-gap",
      "personaId": "USR-SK-03",
      "title": "Replenishment cadence and capability gap",
      "description": "Sofia repeatedly buys beans and cartridges, owns a canister, and has no grinder on file. The run should separate replenishment from discovery and identify the missing capability without recommending another owned durable.",
      "criterionIds": [
        "finds-capability-gap",
        "separates-lanes",
        "avoids-owned-durable",
        "grounds-advice"
      ]
    }
  ],
  "scenarioAcceptances": [],
  "arms": [
    {
      "armId": "discovery-workflow-live",
      "architecture": "workflow",
      "repetitions": 1,
      "checks": [
        {
          "key": "vitrine.live.use-case-quality",
          "name": "Canonical use-case criteria",
          "census": {
            "measured": 1,
            "notApplicable": 0,
            "notMeasured": 0,
            "total": 1
          },
          "reliability": {
            "measurement": "measured",
            "successes": 1,
            "total": 1,
            "estimate": 1.0,
            "lower": 0.20654931437723742,
            "upper": 1.0
          }
        },
        {
          "key": "vitrine.live.response-observed",
          "name": "A non-empty customer response was instrumented",
          "census": {
            "measured": 1,
            "notApplicable": 0,
            "notMeasured": 0,
            "total": 1
          },
          "reliability": {
            "measurement": "measured",
            "successes": 1,
            "total": 1,
            "estimate": 1.0,
            "lower": 0.20654931437723742,
            "upper": 1.0
          }
        },
        {
          "key": "vitrine.live.agent-tool-journal",
          "name": "Agent tool journal is complete and read-only",
          "census": {
            "measured": 0,
            "notApplicable": 1,
            "notMeasured": 0,
            "total": 1
          },
          "reliability": {
            "measurement": "notMeasured",
            "successes": 0,
            "total": 0,
            "estimate": null,
            "lower": null,
            "upper": null
          }
        },
        {
          "key": "vitrine.live.workflow-trace",
          "name": "Workflow executors and routes completed without failure",
          "census": {
            "measured": 1,
            "notApplicable": 0,
            "notMeasured": 0,
            "total": 1
          },
          "reliability": {
            "measurement": "measured",
            "successes": 1,
            "total": 1,
            "estimate": 1.0,
            "lower": 0.20654931437723742,
            "upper": 1.0
          }
        }
      ]
    }
  ],
  "comparisons": [],
  "trials": [
    {
      "scenarioId": "sofia-capability-gap",
      "personaId": "USR-SK-03",
      "armId": "discovery-workflow-live",
      "architecture": "workflow",
      "repetition": 1,
      "measurement": "measured",
      "subjectStatus": "completed",
      "checks": [
        {
          "key": "vitrine.live.use-case-quality",
          "name": "Canonical use-case criteria",
          "measurement": "measured",
          "score": 1.0,
          "passed": true
        },
        {
          "key": "vitrine.live.response-observed",
          "name": "A non-empty customer response was instrumented",
          "measurement": "measured",
          "score": 1.0,
          "passed": true
        },
        {
          "key": "vitrine.live.agent-tool-journal",
          "name": "Agent tool journal is complete and read-only",
          "measurement": "notApplicable"
        },
        {
          "key": "vitrine.live.workflow-trace",
          "name": "Workflow executors and routes completed without failure",
          "measurement": "measured",
          "score": 1.0,
          "passed": true
        }
      ],
      "criteria": [
        {
          "id": "finds-capability-gap",
          "measurement": "measured",
          "met": true
        },
        {
          "id": "separates-lanes",
          "measurement": "measured",
          "met": true
        },
        {
          "id": "avoids-owned-durable",
          "measurement": "measured",
          "met": true
        },
        {
          "id": "grounds-advice",
          "measurement": "measured",
          "met": true
        }
      ],
      "subjectUsage": {
        "status": "measured",
        "modelCalls": 3,
        "inputTokens": 7219,
        "outputTokens": 6908,
        "totalTokens": 14127,
        "estimatedCostUsd": 0.174255
      },
      "judgeUsage": {
        "status": "measured",
        "modelCalls": null,
        "inputTokens": null,
        "outputTokens": null,
        "totalTokens": 1998,
        "estimatedCostUsd": 0.022035000000000003
      },
      "passed": true,
      "failure": null,
      "workflow": {
        "executors": [
          {
            "executorId": "InterestMapper",
            "executionCount": 1
          },
          {
            "executorId": "Discovery",
            "executionCount": 2
          },
          {
            "executorId": "CoverageReviewer",
            "executionCount": 2
          },
          {
            "executorId": "Ranker",
            "executionCount": 1
          },
          {
            "executorId": "Presenter",
            "executionCount": 1
          }
        ],
        "routes": [
          "map-to-discovery",
          "discovery-to-review",
          "review-to-more-discovery",
          "discovery-to-review",
          "review-to-ranker",
          "ranker-to-presenter"
        ],
        "discoveryRounds": 2,
        "maximumRounds": 3,
        "superSteps": 7,
        "stopReason": "GapsUnresolvable",
        "looped": true,
        "failureCount": 0,
        "degradationCount": 0,
        "degradationKinds": [],
        "unknownExecutorCount": 0,
        "unknownRouteCount": 0,
        "providerStages": [
          {
            "executorId": "InterestMapper",
            "attemptCount": 1,
            "responseCount": 1,
            "unusableAttemptCount": 0,
            "failedAttemptCount": 0,
            "cancelledAttemptCount": 0,
            "status": "completed",
            "lastUnusableAttemptNumber": 0,
            "lastUsableResponseAttemptNumber": 1
          },
          {
            "executorId": "Presenter",
            "attemptCount": 1,
            "responseCount": 1,
            "unusableAttemptCount": 0,
            "failedAttemptCount": 0,
            "cancelledAttemptCount": 0,
            "status": "completed",
            "lastUnusableAttemptNumber": 0,
            "lastUsableResponseAttemptNumber": 3
          },
          {
            "executorId": "Ranker",
            "attemptCount": 1,
            "responseCount": 1,
            "unusableAttemptCount": 0,
            "failedAttemptCount": 0,
            "cancelledAttemptCount": 0,
            "status": "completed",
            "lastUnusableAttemptNumber": 0,
            "lastUsableResponseAttemptNumber": 2
          }
        ],
        "providerFailedAttemptCount": 0,
        "recoveredProviderFailedAttemptCount": 0,
        "terminalProviderStageCount": 0
      }
    }
  ],
  "criteriaCensus": [
    {
      "scenarioId": "sofia-capability-gap",
      "armId": "discovery-workflow-live",
      "criterionId": "finds-capability-gap",
      "observed": 1,
      "measured": 1,
      "met": 1,
      "unmet": 0,
      "notApplicable": 0,
      "notMeasured": 0
    },
    {
      "scenarioId": "sofia-capability-gap",
      "armId": "discovery-workflow-live",
      "criterionId": "separates-lanes",
      "observed": 1,
      "measured": 1,
      "met": 1,
      "unmet": 0,
      "notApplicable": 0,
      "notMeasured": 0
    },
    {
      "scenarioId": "sofia-capability-gap",
      "armId": "discovery-workflow-live",
      "criterionId": "avoids-owned-durable",
      "observed": 1,
      "measured": 1,
      "met": 1,
      "unmet": 0,
      "notApplicable": 0,
      "notMeasured": 0
    },
    {
      "scenarioId": "sofia-capability-gap",
      "armId": "discovery-workflow-live",
      "criterionId": "grounds-advice",
      "observed": 1,
      "measured": 1,
      "met": 1,
      "unmet": 0,
      "notApplicable": 0,
      "notMeasured": 0
    }
  ],
  "usageTotals": {
    "subject": {
      "evidenceRows": 1,
      "statusCensus": {
        "notReported": 0,
        "measuredZero": 0,
        "measured": 1,
        "lowerBound": 0
      },
      "reportedModelCallRows": 1,
      "reportedModelCalls": 3,
      "reportedInputTokenRows": 1,
      "reportedInputTokens": 7219,
      "reportedOutputTokenRows": 1,
      "reportedOutputTokens": 6908,
      "reportedTotalTokenRows": 1,
      "reportedTotalTokens": 14127,
      "reportedCostRows": 1,
      "estimatedCostUsd": 0.174255
    },
    "judge": {
      "evidenceRows": 1,
      "statusCensus": {
        "notReported": 0,
        "measuredZero": 0,
        "measured": 1,
        "lowerBound": 0
      },
      "reportedModelCallRows": 0,
      "reportedModelCalls": null,
      "reportedInputTokenRows": 0,
      "reportedInputTokens": null,
      "reportedOutputTokenRows": 0,
      "reportedOutputTokens": null,
      "reportedTotalTokenRows": 1,
      "reportedTotalTokens": 1998,
      "reportedCostRows": 1,
      "estimatedCostUsd": 0.022035000000000003
    }
  },
  "failures": [],
  "safety": null,
  "publicationBoundary": "Allow-listed aggregates only; raw prompts, queries, expected answers, ground truth, model responses, tool arguments, judge explanations, provider errors, endpoints, credentials, canaries, absolute paths and run directories are not published.",
  "passThreshold": 1.0
}