{
  "boundaries": {
    "composite_score": null,
    "external_cost_usd": "13.0413660",
    "leaderboard": false,
    "model_downloads": 0,
    "model_invocations": 3058,
    "network_calls": 3058,
    "performance_results_exist": false,
    "private_material_exists": true,
    "provider_calls": 3058
  },
  "campaign_registration": "fresh-v5-engineering-complete-prior-campaigns-preserved",
  "custody_path_conformance": {
    "algorithm": "AES-256-GCM",
    "classification": "redacted-engineering-only-no-performance-evidence",
    "cost_usd": "0",
    "decision": "custody-root-cause-repaired-live-conformance-pass",
    "exact_cause": "v2-campaign-absent-from-private-generation-permit-allowlist",
    "independent_verifier": "pass",
    "negative_classes": 6,
    "private_read_authority_closed": true,
    "provider_calls": 0,
    "required_classes": 20,
    "required_classes_passed": 20
  },
  "display": {
    "formal_suite": "TreasureBench",
    "historical_alias": "DiscoveryBench",
    "role": "formal-instrument"
  },
  "fixed_batch_diagnostic": {
    "authenticated_fixed_batch_traces": 450,
    "causal_class": "protocol-contract-nonconformance",
    "classification": "redacted-engineering-only-no-performance-evidence",
    "cost_usd": "0",
    "decision": "agent-protocol-policy-decision-required",
    "parse_or_schema_repair_exhaustion_traces": 32,
    "private_read_authority_closed": true,
    "protocol_nonconformance_traces": 32,
    "provider_calls": 0,
    "retained_state_mutated": false
  },
  "fresh_v4_closeout": {
    "batch_id": "tb-agents-v1-repair-confirmation-v4-b01",
    "calls": 116,
    "campaign_id": "treasurebench-agents-v1-repair-confirmation-v4",
    "cost_usd": "0.3470610",
    "decision": "fresh-pilot-v4-quarantined-engineering-only",
    "failure_class": "provider-terminal-missing",
    "input_tokens": 58636,
    "minimum_unseal_performed": false,
    "objects_locked": 150,
    "output_lock_commitment": "sha256:88c442bf6cb5e1b99a893808d423b488d2b1062543a49fcfb32b77f276abaa1c",
    "output_tokens": 13275,
    "performance_results_exist": false,
    "provider_phase_closed": true,
    "real_batch_protocol_classification": "not-run-incomplete-pairings",
    "stage": "private-prefix"
  },
  "fresh_v4_preparation": {
    "architecture_contrast_bounds": 120,
    "batch_id": "tb-agents-v1-repair-confirmation-v4-b01",
    "campaign_id": "treasurebench-agents-v1-repair-confirmation-v4",
    "credential_reads": 0,
    "external_cost_usd": "0",
    "intended_pairings": 500,
    "metric_intervals": 6400,
    "owner_gate_required": false,
    "provider_calls": 0,
    "status": "pass-offline-exact-execution-frozen-owner-gate-required",
    "task_id": "AO-0010",
    "tasks": 50,
    "terminal_pairings_classified": 500,
    "total_relevant_corruptions_rejected": 104,
    "zero_mixed_all_invalid_scenarios": "pass"
  },
  "fresh_v5_closeout": {
    "batch_id": "tb-agents-v1-repair-confirmation-v5-b01",
    "calls": 3058,
    "campaign_id": "treasurebench-agents-v1-repair-confirmation-v5",
    "circuit_breaker_fired": false,
    "cost_usd": "13.0413660",
    "decision": "fresh-pilot-v5-engineering-complete-no-scientific-evidence",
    "independent_metric_bound_agreement": true,
    "input_tokens": 2290993,
    "objects_locked": 4067,
    "output_lock_commitment": "sha256:e18e7f8173f9ac0026f74cd6ff9b577010abdf65620ad269606f6862ff16e47b",
    "output_tokens": 437027,
    "performance_results_exist": false,
    "protocol_invalid_pairings": 4,
    "protocol_valid_pairings": 496,
    "provider_contract_or_safety_failure_pairings": 0,
    "provider_operational_missing_pairings": 0,
    "provider_phase_closed": true,
    "retry_delay_seconds": 5,
    "terminal_pairings_classified": 500,
    "third_transport_attempts": 0,
    "unseal_after_lock_verified": true
  },
  "future_authorization_required": false,
  "generator_counts": {
    "canonical_cells": 138,
    "isomorphic_prompt_variants": 552,
    "primitive_labeled_states": 58945
  },
  "instrument_id": "discoverybench-agents-v1",
  "next_gate": "new explicit owner decision and wholly new identities required for any separate future task",
  "owner": "DD-010",
  "prospective_policy_v2": {
    "global_protocol_conformance_threshold": null,
    "policy_version": "treasurebench-agents-v1-protocol-validity-policy-v2",
    "protocol_invalid_alone_quarantines": false,
    "provider_or_model_ranking": false,
    "status": "prospective-frozen-not-executed",
    "statuses": [
      "batch-integrity-valid",
      "provider-terminal-missing",
      "protocol-valid",
      "protocol-invalid"
    ]
  },
  "provider_schema_conformance": {
    "bisection_calls": 0,
    "calls": 4,
    "classification": "public-engineering-only-not-scientific-evidence",
    "cost_usd": "0.0086685",
    "decision": "conformance-pass-both-complete-schemas",
    "input_tokens": 1603,
    "output_tokens": 275
  },
  "public_rehearsal": {
    "architectures": 5,
    "cases": 50,
    "corruptions_rejected": 24,
    "method_a_b_agree": true,
    "rehearsal_hash": "sha256:d3410ff04bb73dcae929c3abc4cf289d58d6830f2a5ab50ca53764bef4af2c59",
    "tasks": 10
  },
  "repair_rehearsal": {
    "cases": 50,
    "corruptions_rejected": 28,
    "method_a_b_agree": true,
    "method_c_errors": 0,
    "rehearsal_hash": "sha256:d13d925886b96015812ffd79e59faa89e2672a0efe7774652e3436b0c8c70d75",
    "status": "pass"
  },
  "schema_version": 1,
  "status": "implementation-and-custody-path-repaired-five-pilots-quarantined-fresh-v5-engineering-complete-policies-v2-v3-frozen",
  "versions": {
    "agent_protocol": "agents-v1",
    "benchmark_content": [
      "v1",
      "v2",
      "v3"
    ],
    "contamination_report": "agents-contamination-v1",
    "custody_manifest": "agents-custody-v1",
    "prompt_template": "agents-prompt-v1",
    "result": "agents-result-v1",
    "statistical_analysis": "agents-statistical-analysis-v1",
    "structured_action": "agents-action-v1",
    "task_generator": "agents-task-generator-v1",
    "trace": "agents-trace-v1"
  }
}
