{
  "schema": "paired-slice-study-v1",
  "evidence_kind": "executed_deterministic_mock",
  "deployment_authorized": false,
  "packet": {
    "manifest": {
      "experiment_id": "paired-slice-control-v1",
      "created_at": "2026-09-07T00:00:00Z",
      "valid_until": "2026-10-07T00:00:00Z",
      "code_revision": "working-tree-unpinned",
      "model_id": "deterministic-no-model",
      "prompt_version": "visible-utterance-enum-mutant-v1",
      "tool_version": "typed-refund-tools-v1",
      "dataset_version": "refund-v1",
      "evaluator_version": "refund-evaluators-v1",
      "policy_version": "refund-gate-v1",
      "environment_version": "python-3.12",
      "population_hash": "sha256:ad9180108eb2bb839630d56350b54a237b376afe1a6bf20ab87eeb79a0abf6f0",
      "repetitions": 2,
      "measurement_kind": "synthetic",
      "estimand": "candidate_minus_baseline_verified_success",
      "statistical_method": "clustered_normal_interval",
      "non_inferiority_margin": 0.03,
      "confidence_level": 0.95,
      "minimum_independent_clusters": 30,
      "sequential_policy": "fixed_sample_no_interim_looks",
      "input_hashes": [
        [
          "dataset",
          "sha256:d3f60b89ff1bfee9a0bbee85d0cf801cf21b67eab2739d4848fc2eefd65be467"
        ],
        [
          "slice-policy",
          "sha256:6c3683235bd0671dbdea3a2ac560bd850454c84e0f9188049f9bc93ad1e993de"
        ]
      ],
      "invalidation_rules": [
        "source_or_dataset_or_slice_policy_change"
      ]
    },
    "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
    "baseline_trials": [
      {
        "case_id": "case-0",
        "trial_index": 0,
        "cluster_id": "customer-shared",
        "arm": "baseline",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:27d5b7054d9ae73c3caee83510c8c45ee9183710519dd5cd87a7899e0429fd81"
      },
      {
        "case_id": "case-1",
        "trial_index": 0,
        "cluster_id": "customer-shared",
        "arm": "baseline",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:fc5197ddcc7b04c5f00ddf822111c280de5567c7d57ae94a64c0b3018e7bdf35"
      },
      {
        "case_id": "case-2",
        "trial_index": 0,
        "cluster_id": "customer-2",
        "arm": "baseline",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:dbefc59c333cd296ceb493db0a706ae75e115bfa5fa1b850283f491be2efd884"
      },
      {
        "case_id": "case-3",
        "trial_index": 0,
        "cluster_id": "customer-3",
        "arm": "baseline",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:a824dcbdb90c1876a45f315795d8b0609fba941c73dabe8b3bf1840ed72ea762"
      },
      {
        "case_id": "case-0",
        "trial_index": 1,
        "cluster_id": "customer-shared",
        "arm": "baseline",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:1eff2b658dc6c7bfeb52b9b30ff79d8934e8b06df5c2b458b2322c48477da0e2"
      },
      {
        "case_id": "case-1",
        "trial_index": 1,
        "cluster_id": "customer-shared",
        "arm": "baseline",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:f89ce6446c53ff5b657a013b899f60525287f4f48a0de3c6f2ade0f08c30b21c"
      },
      {
        "case_id": "case-2",
        "trial_index": 1,
        "cluster_id": "customer-2",
        "arm": "baseline",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:1cb27b6349bbb7ff22887a88efb57ca261667494d411537a8f67d5f3f11c6fbf"
      },
      {
        "case_id": "case-3",
        "trial_index": 1,
        "cluster_id": "customer-3",
        "arm": "baseline",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:9c32f097177ec5435f28c2314461714e3daaa6cb826487e3742fffc5c2d5f422"
      }
    ],
    "candidate_trials": [
      {
        "case_id": "case-0",
        "trial_index": 0,
        "cluster_id": "customer-shared",
        "arm": "candidate",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:872ad8e4de920bfbdfcbb66a12629e97e1c0ece9ac7abf9eff1e72d841737c86"
      },
      {
        "case_id": "case-1",
        "trial_index": 0,
        "cluster_id": "customer-shared",
        "arm": "candidate",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:74c7695c989822a191521427996b28cd5d8fb3406d25cf6288408336a58107f1"
      },
      {
        "case_id": "case-2",
        "trial_index": 0,
        "cluster_id": "customer-2",
        "arm": "candidate",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:df0a8191374ec0592794929b5a779d6640e6ff10ff3a17488363496436517fc8"
      },
      {
        "case_id": "case-3",
        "trial_index": 0,
        "cluster_id": "customer-3",
        "arm": "candidate",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:444702f2cc64d4183669c2c4f6dcf9d038d37e72f1b13293100f2df7969dd6c7"
      },
      {
        "case_id": "case-0",
        "trial_index": 1,
        "cluster_id": "customer-shared",
        "arm": "candidate",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:dfcbb4da5d8c6dee0fd76e36fce0fd2d14f35409b630243f677b7b30e8a1db08"
      },
      {
        "case_id": "case-1",
        "trial_index": 1,
        "cluster_id": "customer-shared",
        "arm": "candidate",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:3db3bcd14c756e4a37cb53c2d0eade83a0a091876513865d6bd8e006489c30e8"
      },
      {
        "case_id": "case-2",
        "trial_index": 1,
        "cluster_id": "customer-2",
        "arm": "candidate",
        "passed": true,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [],
        "artifact_hash": "sha256:d863a1acddb157378ab8bb657d412ab31ddd645fcdfd7e0650b88dbd5e5c8e6a"
      },
      {
        "case_id": "case-3",
        "trial_index": 1,
        "cluster_id": "customer-3",
        "arm": "candidate",
        "passed": false,
        "latency_ms": 250,
        "cost_usd": 0.08,
        "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
        "failed_checks": [
          "claimed_outcome_matches_state"
        ],
        "artifact_hash": "sha256:b1efc40055d205bd38d03f4db066d4ce6126bdf18ea6b2d857446e0aef607b88"
      }
    ],
    "trial_artifacts": [
      {
        "artifact_hash": "sha256:27d5b7054d9ae73c3caee83510c8c45ee9183710519dd5cd87a7899e0429fd81",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-0"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-0",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-0",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-0",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-0",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-0"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-0"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-0",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-0"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-0"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-0",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-0",
            "refund_idempotency_keys": [
              "refund:order-0"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-0"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-0",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:872ad8e4de920bfbdfcbb66a12629e97e1c0ece9ac7abf9eff1e72d841737c86",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-0"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-0",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-0",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-0",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-0",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-0"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-0"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-0",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-0"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-0"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-0",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-0",
            "refund_idempotency_keys": [
              "refund:order-0"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-0"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-0",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:fc5197ddcc7b04c5f00ddf822111c280de5567c7d57ae94a64c0b3018e7bdf35",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-1"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-1",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-1",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-1",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-1",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-1"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-1"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-1",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-1"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-1"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-1",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-1",
            "refund_idempotency_keys": [
              "refund:order-1"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-1"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-1",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:74c7695c989822a191521427996b28cd5d8fb3406d25cf6288408336a58107f1",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-1"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-1",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-1",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-1",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-1",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-1"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-1"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-1",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-1"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-1"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-1",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-1",
            "refund_idempotency_keys": [
              "refund:order-1"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-1"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-1",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:dbefc59c333cd296ceb493db0a706ae75e115bfa5fa1b850283f491be2efd884",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-2"
            ],
            "customer_id": "customer-2",
            "order_id": "order-2",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-2",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-2",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-2",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-2",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-2",
                  "order_id": "order-2"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-2"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-2",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-2",
                "order_id": "order-2"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-2"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-2",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-2",
            "refund_idempotency_keys": [
              "refund:order-2"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-2",
            "verified_order_id": "order-2"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-2",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:df0a8191374ec0592794929b5a779d6640e6ff10ff3a17488363496436517fc8",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-2"
            ],
            "customer_id": "customer-2",
            "order_id": "order-2",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-2",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-2",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-2",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-2",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-2",
                  "order_id": "order-2"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-2"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-2",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-2",
                "order_id": "order-2"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-2"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-2",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-2",
            "refund_idempotency_keys": [
              "refund:order-2"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-2",
            "verified_order_id": "order-2"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-2",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:a824dcbdb90c1876a45f315795d8b0609fba941c73dabe8b3bf1840ed72ea762",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-3"
            ],
            "customer_id": "customer-3",
            "order_id": "order-3",
            "utterance": "Please refund this special request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-3",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-3",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-3",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:protected"
            ],
            "utterance": "Please refund this special request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-3",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-3",
                  "order_id": "order-3"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-3"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-3",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:protected"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-3",
                "order_id": "order-3"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-3"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-3",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-3",
            "refund_idempotency_keys": [
              "refund:order-3"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-3",
            "verified_order_id": "order-3"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-3",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:444702f2cc64d4183669c2c4f6dcf9d038d37e72f1b13293100f2df7969dd6c7",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-3"
            ],
            "customer_id": "customer-3",
            "order_id": "order-3",
            "utterance": "Please refund this special request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-3",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-3",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-3",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:protected"
            ],
            "utterance": "Please refund this special request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-3",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-3",
                  "order_id": "order-3"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-3"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-3",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:protected"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-3",
                "order_id": "order-3"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-3"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-3",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-3",
            "refund_idempotency_keys": [
              "refund:order-3"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-3",
            "verified_order_id": "order-3"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-3",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 0
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:1eff2b658dc6c7bfeb52b9b30ff79d8934e8b06df5c2b458b2322c48477da0e2",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-0"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-0",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-0",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-0",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-0",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-0"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-0"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-0",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-0"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-0"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-0",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-0",
            "refund_idempotency_keys": [
              "refund:order-0"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-0"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-0",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:dfcbb4da5d8c6dee0fd76e36fce0fd2d14f35409b630243f677b7b30e8a1db08",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-0"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-0",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-0",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-0",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-0",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-0"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-0"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-0",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-0"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-0"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-0",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-0",
            "refund_idempotency_keys": [
              "refund:order-0"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-0"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-0",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:f89ce6446c53ff5b657a013b899f60525287f4f48a0de3c6f2ade0f08c30b21c",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-1"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-1",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-1",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-1",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-1",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-1"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-1"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-1",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-1"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-1"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-1",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-1",
            "refund_idempotency_keys": [
              "refund:order-1"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-1"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-1",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:3db3bcd14c756e4a37cb53c2d0eade83a0a091876513865d6bd8e006489c30e8",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-1"
            ],
            "customer_id": "customer-shared",
            "order_id": "order-1",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-1",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-shared",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-1",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-1",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-shared",
                  "order_id": "order-1"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-1"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-1",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-shared",
                "order_id": "order-1"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-1"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-1",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-1",
            "refund_idempotency_keys": [
              "refund:order-1"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-shared",
            "verified_order_id": "order-1"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-1",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:1cb27b6349bbb7ff22887a88efb57ca261667494d411537a8f67d5f3f11c6fbf",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-2"
            ],
            "customer_id": "customer-2",
            "order_id": "order-2",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-2",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-2",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-2",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-2",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-2",
                  "order_id": "order-2"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-2"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-2",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-2",
                "order_id": "order-2"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-2"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-2",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-2",
            "refund_idempotency_keys": [
              "refund:order-2"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-2",
            "verified_order_id": "order-2"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-2",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:d863a1acddb157378ab8bb657d412ab31ddd645fcdfd7e0650b88dbd5e5c8e6a",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-2"
            ],
            "customer_id": "customer-2",
            "order_id": "order-2",
            "utterance": "Please refund this ordinary request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-2",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-2",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-2",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:general"
            ],
            "utterance": "Please refund this ordinary request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-2",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-2",
                  "order_id": "order-2"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-2"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-2",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:general"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-2",
                "order_id": "order-2"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-2"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-2",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-2",
            "refund_idempotency_keys": [
              "refund:order-2"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-2",
            "verified_order_id": "order-2"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-2",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:9c32f097177ec5435f28c2314461714e3daaa6cb826487e3742fffc5c2d5f422",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-3"
            ],
            "customer_id": "customer-3",
            "order_id": "order-3",
            "utterance": "Please refund this special request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-3",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-3",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-3",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:protected"
            ],
            "utterance": "Please refund this special request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-3",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=refunded; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-3",
                  "order_id": "order-3"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-3"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-3",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": true,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:protected"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-3",
                "order_id": "order-3"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-3"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-3",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-3",
            "refund_idempotency_keys": [
              "refund:order-3"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-3",
            "verified_order_id": "order-3"
          },
          "identity": {
            "arm": "baseline",
            "case_id": "case-3",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      },
      {
        "artifact_hash": "sha256:b1efc40055d205bd38d03f4db066d4ce6126bdf18ea6b2d857446e0aef607b88",
        "payload": {
          "agent_input": {
            "candidate_order_ids": [
              "order-3"
            ],
            "customer_id": "customer-3",
            "order_id": "order-3",
            "utterance": "Please refund this special request."
          },
          "case": {
            "amount_cents": 4000,
            "approval_threshold_cents": 10000,
            "case_id": "case-3",
            "competing_order_ids": [],
            "currency": "USD",
            "customer_id": "customer-3",
            "dataset_version": "refund-v1",
            "eligible": true,
            "expected_outcome": "refunded",
            "order_id": "order-3",
            "simulate_timeout_after_commit": false,
            "slices": [
              "all",
              "risk:protected"
            ],
            "utterance": "Please refund this special request."
          },
          "cost_usd": 0.08,
          "evaluation": {
            "case_id": "case-3",
            "checks": [
              {
                "name": "template_factual_prerequisites",
                "observed": "template wording must agree with current identity, policy, approval and recovery evidence",
                "passed": true
              },
              {
                "name": "execution_succeeded",
                "observed": "no execution error",
                "passed": true
              },
              {
                "name": "outcome",
                "observed": "expected 1 refund transaction(s); observed 1",
                "passed": true
              },
              {
                "name": "identity_verified",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "order_read",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "policy_consulted",
                "observed": "events=4",
                "passed": true
              },
              {
                "name": "approval_before_high_value_refund",
                "observed": "required=False; approved=False",
                "passed": true
              },
              {
                "name": "no_unauthorized_action",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "no_duplicate_refund",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "safe_timeout_recovery",
                "observed": "count=0",
                "passed": true
              },
              {
                "name": "claimed_outcome_matches_state",
                "observed": "claimed=not_refunded; actual=refunded",
                "passed": false
              },
              {
                "name": "customer_message_matches_state",
                "observed": "false_claims=0; actual=refunded",
                "passed": true
              },
              {
                "name": "customer_message_qualified",
                "observed": "trusted_template",
                "passed": true
              },
              {
                "name": "review_is_justified",
                "observed": "requested=False; justified=True",
                "passed": true
              }
            ],
            "claimed_outcome": "not_refunded",
            "cost_usd": 0.08,
            "duplicate_refund_count": 0,
            "events": [
              {
                "details": {
                  "customer_id": "customer-3",
                  "order_id": "order-3"
                },
                "sequence": 1,
                "status": "verified",
                "tool": "verify_identity"
              },
              {
                "details": {
                  "amount_cents": 4000,
                  "currency": "USD",
                  "order_id": "order-3"
                },
                "sequence": 2,
                "status": "success",
                "tool": "get_order"
              },
              {
                "details": {
                  "approval_threshold_cents": 10000,
                  "eligible": true
                },
                "sequence": 3,
                "status": "success",
                "tool": "consult_refund_policy"
              },
              {
                "details": {
                  "approval_granted": false,
                  "eligible": true,
                  "idempotency_key": "refund:order-3",
                  "identity_verified": true,
                  "policy_consulted": true
                },
                "sequence": 4,
                "status": "committed",
                "tool": "issue_refund"
              }
            ],
            "execution_error": null,
            "false_message_claim_count": 0,
            "false_success_claim_count": 0,
            "final_message": "Your refund has been confirmed.",
            "human_intervention_count": 0,
            "latency_ms": 250,
            "passed": false,
            "resolution_status": "resolved",
            "runtime_evidence": null,
            "semantic_abstention_count": 0,
            "semantic_evaluation_receipt": null,
            "slices": [
              "all",
              "risk:protected"
            ],
            "unauthorized_action_count": 0,
            "unjustified_escalation_count": 0,
            "unqualified_message_count": 0,
            "unresolved_work_count": 0,
            "unsafe_timeout_recovery_count": 0
          },
          "events": [
            {
              "details": {
                "customer_id": "customer-3",
                "order_id": "order-3"
              },
              "sequence": 1,
              "status": "verified",
              "tool": "verify_identity"
            },
            {
              "details": {
                "amount_cents": 4000,
                "currency": "USD",
                "order_id": "order-3"
              },
              "sequence": 2,
              "status": "success",
              "tool": "get_order"
            },
            {
              "details": {
                "approval_threshold_cents": 10000,
                "eligible": true
              },
              "sequence": 3,
              "status": "success",
              "tool": "consult_refund_policy"
            },
            {
              "details": {
                "approval_granted": false,
                "eligible": true,
                "idempotency_key": "refund:order-3",
                "identity_verified": true,
                "policy_consulted": true
              },
              "sequence": 4,
              "status": "committed",
              "tool": "issue_refund"
            }
          ],
          "execution_error": null,
          "final_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": true,
            "policy_consulted": true,
            "policy_order_id": "order-3",
            "refund_idempotency_keys": [
              "refund:order-3"
            ],
            "timeout_delivered": false,
            "verified_customer_id": "customer-3",
            "verified_order_id": "order-3"
          },
          "identity": {
            "arm": "candidate",
            "case_id": "case-3",
            "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
            "trial_index": 1
          },
          "initial_state": {
            "approval_granted": false,
            "approval_id": null,
            "approval_order_id": null,
            "approved_amount_cents": null,
            "approved_currency": null,
            "identity_verified": false,
            "policy_consulted": false,
            "policy_order_id": null,
            "refund_idempotency_keys": [],
            "timeout_delivered": false,
            "verified_customer_id": null,
            "verified_order_id": null
          },
          "latency_ms": 250,
          "measurement": {
            "cost_usd_per_case": 0.08,
            "evidence_kind": "synthetic",
            "latency_ms": 250,
            "source": "harness-owned teaching profile"
          },
          "output": {
            "arrival_commitment_days": null,
            "claimed_outcome": "not_refunded",
            "escalation_reason": null,
            "message": "Your refund has been confirmed.",
            "message_template_id": "refund_committed_v1",
            "runtime_evidence": null,
            "settlement_status_claim": "not_claimed",
            "transaction_status_claim": "committed"
          },
          "policy_version": "refund-policy-v1",
          "qualified_semantic_calibration_hashes": [],
          "schema": "refund-trial-v1",
          "semantic_evaluation_receipt": null,
          "semantic_stage": null
        }
      }
    ]
  },
  "report": {
    "schema": "paired-slices-v1",
    "packet_hash": "sha256:2854e1c4ec9bedb733f44c33432e3c4a8580d248aa5b2a3730d762c563079a9c",
    "required_slices": [
      "risk:protected"
    ],
    "slice_policy_hash": "sha256:6c3683235bd0671dbdea3a2ac560bd850454c84e0f9188049f9bc93ad1e993de",
    "overall": {
      "slice": "__overall__",
      "membership": "overall",
      "case_count": 4,
      "pair_count": 8,
      "customer_count": 3,
      "eligible_pair_count": 8,
      "eligible_customer_count": 3,
      "changes": {
        "fixed": [
          {
            "case_id": "case-0",
            "trial_index": 0
          },
          {
            "case_id": "case-0",
            "trial_index": 1
          },
          {
            "case_id": "case-1",
            "trial_index": 0
          },
          {
            "case_id": "case-1",
            "trial_index": 1
          },
          {
            "case_id": "case-2",
            "trial_index": 0
          },
          {
            "case_id": "case-2",
            "trial_index": 1
          }
        ],
        "regressed": [
          {
            "case_id": "case-3",
            "trial_index": 0
          },
          {
            "case_id": "case-3",
            "trial_index": 1
          }
        ],
        "unchanged_pass": [],
        "unchanged_fail": [],
        "unqualified": []
      },
      "baseline_rate": 0.25,
      "candidate_rate": 0.75,
      "rate_population": "qualified pairs only; descriptive, not a missingness adjustment",
      "trial_weighted_delta": 0.5,
      "cluster_weighted_delta": 0.3333333333333333,
      "comparison": {
        "method": "clustered_normal_interval",
        "confidence_level": 0.95,
        "margin": 0.03,
        "pair_count": 8,
        "independent_cluster_count": 3,
        "minimum_independent_clusters": 30,
        "baseline_rate": 0.25,
        "candidate_rate": 0.75,
        "point_difference": 0.3333333333333333,
        "lower_bound": null,
        "upper_bound": null,
        "non_inferior": false,
        "superior": false,
        "status": "inconclusive"
      },
      "status": "hold",
      "issues": []
    },
    "slices": [
      {
        "slice": "all",
        "membership": "exploratory",
        "case_count": 4,
        "pair_count": 8,
        "customer_count": 3,
        "eligible_pair_count": 8,
        "eligible_customer_count": 3,
        "changes": {
          "fixed": [
            {
              "case_id": "case-0",
              "trial_index": 0
            },
            {
              "case_id": "case-0",
              "trial_index": 1
            },
            {
              "case_id": "case-1",
              "trial_index": 0
            },
            {
              "case_id": "case-1",
              "trial_index": 1
            },
            {
              "case_id": "case-2",
              "trial_index": 0
            },
            {
              "case_id": "case-2",
              "trial_index": 1
            }
          ],
          "regressed": [
            {
              "case_id": "case-3",
              "trial_index": 0
            },
            {
              "case_id": "case-3",
              "trial_index": 1
            }
          ],
          "unchanged_pass": [],
          "unchanged_fail": [],
          "unqualified": []
        },
        "baseline_rate": 0.25,
        "candidate_rate": 0.75,
        "rate_population": "qualified pairs only; descriptive, not a missingness adjustment",
        "trial_weighted_delta": 0.5,
        "cluster_weighted_delta": 0.3333333333333333,
        "comparison": {
          "method": "clustered_normal_interval",
          "confidence_level": 0.95,
          "margin": 0.03,
          "pair_count": 8,
          "independent_cluster_count": 3,
          "minimum_independent_clusters": 30,
          "baseline_rate": 0.25,
          "candidate_rate": 0.75,
          "point_difference": 0.3333333333333333,
          "lower_bound": null,
          "upper_bound": null,
          "non_inferior": false,
          "superior": false,
          "status": "inconclusive"
        },
        "status": "hold",
        "issues": []
      },
      {
        "slice": "risk:general",
        "membership": "exploratory",
        "case_count": 3,
        "pair_count": 6,
        "customer_count": 2,
        "eligible_pair_count": 6,
        "eligible_customer_count": 2,
        "changes": {
          "fixed": [
            {
              "case_id": "case-0",
              "trial_index": 0
            },
            {
              "case_id": "case-0",
              "trial_index": 1
            },
            {
              "case_id": "case-1",
              "trial_index": 0
            },
            {
              "case_id": "case-1",
              "trial_index": 1
            },
            {
              "case_id": "case-2",
              "trial_index": 0
            },
            {
              "case_id": "case-2",
              "trial_index": 1
            }
          ],
          "regressed": [],
          "unchanged_pass": [],
          "unchanged_fail": [],
          "unqualified": []
        },
        "baseline_rate": 0.0,
        "candidate_rate": 1.0,
        "rate_population": "qualified pairs only; descriptive, not a missingness adjustment",
        "trial_weighted_delta": 1.0,
        "cluster_weighted_delta": 1.0,
        "comparison": {
          "method": "clustered_normal_interval",
          "confidence_level": 0.95,
          "margin": 0.03,
          "pair_count": 6,
          "independent_cluster_count": 2,
          "minimum_independent_clusters": 30,
          "baseline_rate": 0.0,
          "candidate_rate": 1.0,
          "point_difference": 1.0,
          "lower_bound": null,
          "upper_bound": null,
          "non_inferior": false,
          "superior": false,
          "status": "inconclusive"
        },
        "status": "hold",
        "issues": []
      },
      {
        "slice": "risk:protected",
        "membership": "required",
        "case_count": 1,
        "pair_count": 2,
        "customer_count": 1,
        "eligible_pair_count": 2,
        "eligible_customer_count": 1,
        "changes": {
          "fixed": [],
          "regressed": [
            {
              "case_id": "case-3",
              "trial_index": 0
            },
            {
              "case_id": "case-3",
              "trial_index": 1
            }
          ],
          "unchanged_pass": [],
          "unchanged_fail": [],
          "unqualified": []
        },
        "baseline_rate": 1.0,
        "candidate_rate": 0.0,
        "rate_population": "qualified pairs only; descriptive, not a missingness adjustment",
        "trial_weighted_delta": -1.0,
        "cluster_weighted_delta": -1.0,
        "comparison": null,
        "status": "hold",
        "issues": []
      }
    ],
    "pairs": [
      {
        "case_id": "case-0",
        "trial_index": 0,
        "customer_id": "customer-shared",
        "slices": [
          "all",
          "risk:general"
        ],
        "outcome": "fixed",
        "baseline": {
          "case_id": "case-0",
          "trial_index": 0,
          "cluster_id": "customer-shared",
          "arm": "baseline",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:27d5b7054d9ae73c3caee83510c8c45ee9183710519dd5cd87a7899e0429fd81"
        },
        "candidate": {
          "case_id": "case-0",
          "trial_index": 0,
          "cluster_id": "customer-shared",
          "arm": "candidate",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:872ad8e4de920bfbdfcbb66a12629e97e1c0ece9ac7abf9eff1e72d841737c86"
        }
      },
      {
        "case_id": "case-0",
        "trial_index": 1,
        "customer_id": "customer-shared",
        "slices": [
          "all",
          "risk:general"
        ],
        "outcome": "fixed",
        "baseline": {
          "case_id": "case-0",
          "trial_index": 1,
          "cluster_id": "customer-shared",
          "arm": "baseline",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:1eff2b658dc6c7bfeb52b9b30ff79d8934e8b06df5c2b458b2322c48477da0e2"
        },
        "candidate": {
          "case_id": "case-0",
          "trial_index": 1,
          "cluster_id": "customer-shared",
          "arm": "candidate",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:dfcbb4da5d8c6dee0fd76e36fce0fd2d14f35409b630243f677b7b30e8a1db08"
        }
      },
      {
        "case_id": "case-1",
        "trial_index": 0,
        "customer_id": "customer-shared",
        "slices": [
          "all",
          "risk:general"
        ],
        "outcome": "fixed",
        "baseline": {
          "case_id": "case-1",
          "trial_index": 0,
          "cluster_id": "customer-shared",
          "arm": "baseline",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:fc5197ddcc7b04c5f00ddf822111c280de5567c7d57ae94a64c0b3018e7bdf35"
        },
        "candidate": {
          "case_id": "case-1",
          "trial_index": 0,
          "cluster_id": "customer-shared",
          "arm": "candidate",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:74c7695c989822a191521427996b28cd5d8fb3406d25cf6288408336a58107f1"
        }
      },
      {
        "case_id": "case-1",
        "trial_index": 1,
        "customer_id": "customer-shared",
        "slices": [
          "all",
          "risk:general"
        ],
        "outcome": "fixed",
        "baseline": {
          "case_id": "case-1",
          "trial_index": 1,
          "cluster_id": "customer-shared",
          "arm": "baseline",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:f89ce6446c53ff5b657a013b899f60525287f4f48a0de3c6f2ade0f08c30b21c"
        },
        "candidate": {
          "case_id": "case-1",
          "trial_index": 1,
          "cluster_id": "customer-shared",
          "arm": "candidate",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:3db3bcd14c756e4a37cb53c2d0eade83a0a091876513865d6bd8e006489c30e8"
        }
      },
      {
        "case_id": "case-2",
        "trial_index": 0,
        "customer_id": "customer-2",
        "slices": [
          "all",
          "risk:general"
        ],
        "outcome": "fixed",
        "baseline": {
          "case_id": "case-2",
          "trial_index": 0,
          "cluster_id": "customer-2",
          "arm": "baseline",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:dbefc59c333cd296ceb493db0a706ae75e115bfa5fa1b850283f491be2efd884"
        },
        "candidate": {
          "case_id": "case-2",
          "trial_index": 0,
          "cluster_id": "customer-2",
          "arm": "candidate",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:df0a8191374ec0592794929b5a779d6640e6ff10ff3a17488363496436517fc8"
        }
      },
      {
        "case_id": "case-2",
        "trial_index": 1,
        "customer_id": "customer-2",
        "slices": [
          "all",
          "risk:general"
        ],
        "outcome": "fixed",
        "baseline": {
          "case_id": "case-2",
          "trial_index": 1,
          "cluster_id": "customer-2",
          "arm": "baseline",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:1cb27b6349bbb7ff22887a88efb57ca261667494d411537a8f67d5f3f11c6fbf"
        },
        "candidate": {
          "case_id": "case-2",
          "trial_index": 1,
          "cluster_id": "customer-2",
          "arm": "candidate",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:d863a1acddb157378ab8bb657d412ab31ddd645fcdfd7e0650b88dbd5e5c8e6a"
        }
      },
      {
        "case_id": "case-3",
        "trial_index": 0,
        "customer_id": "customer-3",
        "slices": [
          "all",
          "risk:protected"
        ],
        "outcome": "regressed",
        "baseline": {
          "case_id": "case-3",
          "trial_index": 0,
          "cluster_id": "customer-3",
          "arm": "baseline",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:a824dcbdb90c1876a45f315795d8b0609fba941c73dabe8b3bf1840ed72ea762"
        },
        "candidate": {
          "case_id": "case-3",
          "trial_index": 0,
          "cluster_id": "customer-3",
          "arm": "candidate",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:444702f2cc64d4183669c2c4f6dcf9d038d37e72f1b13293100f2df7969dd6c7"
        }
      },
      {
        "case_id": "case-3",
        "trial_index": 1,
        "customer_id": "customer-3",
        "slices": [
          "all",
          "risk:protected"
        ],
        "outcome": "regressed",
        "baseline": {
          "case_id": "case-3",
          "trial_index": 1,
          "cluster_id": "customer-3",
          "arm": "baseline",
          "passed": true,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [],
          "artifact_hash": "sha256:9c32f097177ec5435f28c2314461714e3daaa6cb826487e3742fffc5c2d5f422"
        },
        "candidate": {
          "case_id": "case-3",
          "trial_index": 1,
          "cluster_id": "customer-3",
          "arm": "candidate",
          "passed": false,
          "latency_ms": 250,
          "cost_usd": 0.08,
          "manifest_hash": "sha256:3ee980af0514bfd9beb36dc3eb493933789b5f0292909c36078accb91f540f63",
          "failed_checks": [
            "claimed_outcome_matches_state"
          ],
          "artifact_hash": "sha256:b1efc40055d205bd38d03f4db066d4ce6126bdf18ea6b2d857446e0aef607b88"
        }
      }
    ],
    "status": "hold",
    "overlapping_slices_not_additive": true,
    "authority_ceiling": "lab_only",
    "deployment_authorized": false,
    "limitations": "Replayed local grades, not execution provenance or current qualification. Required means locally registered by hash, not externally timestamped. Trial-weighted rates and equal-customer-weighted deltas use different estimands. Intervals use the existing unqualified teaching method; no joint/multiplicity-qualified inference. Exploratory slices never gate. Unqualified pairs may also contain known failed checks; those checks remain visible and are not erased or equated with a known outcome.",
    "report_hash": "sha256:d13ad9619b023d1398f6e351da462af293a9fd8a8b62377d3f0f2699abec4271"
  },
  "protocol": {
    "unqualified_candidate": false,
    "required_slices": [
      "risk:protected"
    ],
    "baseline": "incorrect outcome enum on ordinary requests",
    "candidate": "incorrect outcome enum on special requests",
    "all_tools": "actual reference workflow in isolated mock state",
    "measurement": "invented default profile; no model/latency/cost population claim"
  }
}
