{
  "status": "PASS",
  "baseline": {
    "id": "conventional-direct-readback-v1",
    "method": "standalone deterministic expected-state checks; no TESVEN verifier-engine imports"
  },
  "corpus": "mvp0-2026-09-28.2",
  "corpus_sha256": "7f7126a8773d933236055f489857b71d37a0a49b28714dc4ca9525666a91453f",
  "input": {
    "tesven_result": "tesven-mvp0-2026-09-28.2-reproduced-2026-10-04.json",
    "tesven_status_counts": {
      "VERIFIED_PASS": 40,
      "VERIFIED_FAIL": 25,
      "INSUFFICIENT_EVIDENCE": 20,
      "HUMAN_REVIEW": 8,
      "NOT_APPLICABLE": 7
    },
    "postgres": "PostgreSQL 18.4 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0, 64-bit"
  },
  "evaluated_cases": 100,
  "case_statuses": {
    "browser-001": "VERIFIED_PASS",
    "browser-002": "VERIFIED_PASS",
    "browser-003": "VERIFIED_PASS",
    "browser-004": "VERIFIED_PASS",
    "browser-005": "VERIFIED_PASS",
    "browser-006": "VERIFIED_PASS",
    "browser-007": "VERIFIED_PASS",
    "browser-008": "VERIFIED_PASS",
    "browser-009": "VERIFIED_PASS",
    "browser-010": "VERIFIED_PASS",
    "browser-011": "VERIFIED_PASS",
    "browser-012": "VERIFIED_PASS",
    "browser-013": "VERIFIED_FAIL",
    "browser-014": "VERIFIED_FAIL",
    "browser-015": "VERIFIED_FAIL",
    "browser-016": "VERIFIED_FAIL",
    "browser-017": "VERIFIED_FAIL",
    "browser-018": "VERIFIED_FAIL",
    "browser-019": "VERIFIED_FAIL",
    "browser-020": "VERIFIED_FAIL",
    "browser-021": "INSUFFICIENT_EVIDENCE",
    "browser-022": "INSUFFICIENT_EVIDENCE",
    "browser-023": "INSUFFICIENT_EVIDENCE",
    "browser-024": "INSUFFICIENT_EVIDENCE",
    "browser-025": "INSUFFICIENT_EVIDENCE",
    "browser-026": "INSUFFICIENT_EVIDENCE",
    "browser-027": "HUMAN_REVIEW",
    "browser-028": "HUMAN_REVIEW",
    "browser-029": "NOT_APPLICABLE",
    "browser-030": "NOT_APPLICABLE",
    "desktop-001": "VERIFIED_PASS",
    "desktop-002": "VERIFIED_PASS",
    "desktop-003": "VERIFIED_PASS",
    "desktop-004": "VERIFIED_PASS",
    "desktop-005": "VERIFIED_PASS",
    "desktop-006": "VERIFIED_PASS",
    "desktop-007": "VERIFIED_PASS",
    "desktop-008": "VERIFIED_PASS",
    "desktop-009": "VERIFIED_FAIL",
    "desktop-010": "VERIFIED_FAIL",
    "desktop-011": "VERIFIED_FAIL",
    "desktop-012": "VERIFIED_FAIL",
    "desktop-013": "VERIFIED_FAIL",
    "desktop-014": "INSUFFICIENT_EVIDENCE",
    "desktop-015": "INSUFFICIENT_EVIDENCE",
    "desktop-016": "INSUFFICIENT_EVIDENCE",
    "desktop-017": "INSUFFICIENT_EVIDENCE",
    "desktop-018": "HUMAN_REVIEW",
    "desktop-019": "HUMAN_REVIEW",
    "desktop-020": "NOT_APPLICABLE",
    "file-001": "VERIFIED_PASS",
    "file-002": "VERIFIED_PASS",
    "file-003": "VERIFIED_PASS",
    "file-004": "VERIFIED_PASS",
    "file-005": "VERIFIED_PASS",
    "file-006": "VERIFIED_PASS",
    "file-007": "VERIFIED_PASS",
    "file-008": "VERIFIED_PASS",
    "file-009": "VERIFIED_PASS",
    "file-010": "VERIFIED_PASS",
    "file-011": "VERIFIED_PASS",
    "file-012": "VERIFIED_PASS",
    "file-013": "VERIFIED_FAIL",
    "file-014": "VERIFIED_FAIL",
    "file-015": "VERIFIED_FAIL",
    "file-016": "VERIFIED_FAIL",
    "file-017": "VERIFIED_FAIL",
    "file-018": "VERIFIED_FAIL",
    "file-019": "VERIFIED_FAIL",
    "file-020": "VERIFIED_FAIL",
    "file-021": "INSUFFICIENT_EVIDENCE",
    "file-022": "INSUFFICIENT_EVIDENCE",
    "file-023": "INSUFFICIENT_EVIDENCE",
    "file-024": "INSUFFICIENT_EVIDENCE",
    "file-025": "INSUFFICIENT_EVIDENCE",
    "file-026": "INSUFFICIENT_EVIDENCE",
    "file-027": "HUMAN_REVIEW",
    "file-028": "HUMAN_REVIEW",
    "file-029": "NOT_APPLICABLE",
    "file-030": "NOT_APPLICABLE",
    "permissions-001": "VERIFIED_PASS",
    "permissions-002": "VERIFIED_PASS",
    "permissions-003": "VERIFIED_PASS",
    "permissions-004": "VERIFIED_PASS",
    "permissions-005": "VERIFIED_PASS",
    "permissions-006": "VERIFIED_PASS",
    "permissions-007": "VERIFIED_PASS",
    "permissions-008": "VERIFIED_PASS",
    "permissions-009": "VERIFIED_FAIL",
    "permissions-010": "VERIFIED_FAIL",
    "permissions-011": "VERIFIED_FAIL",
    "permissions-012": "VERIFIED_FAIL",
    "permissions-013": "INSUFFICIENT_EVIDENCE",
    "permissions-014": "INSUFFICIENT_EVIDENCE",
    "permissions-015": "INSUFFICIENT_EVIDENCE",
    "permissions-016": "INSUFFICIENT_EVIDENCE",
    "permissions-017": "HUMAN_REVIEW",
    "permissions-018": "HUMAN_REVIEW",
    "permissions-019": "NOT_APPLICABLE",
    "permissions-020": "NOT_APPLICABLE"
  },
  "baseline_status_counts": {
    "VERIFIED_PASS": 40,
    "VERIFIED_FAIL": 25,
    "INSUFFICIENT_EVIDENCE": 20,
    "HUMAN_REVIEW": 8,
    "NOT_APPLICABLE": 7
  },
  "comparison": {
    "baseline_casewise_matches_frozen_oracle": {
      "numerator": 100,
      "denominator": 100
    },
    "status_distribution_matches_tesven": true,
    "exact_tesven_casewise_statuses_in_input_artifact": false,
    "same_oracle_transitivity": true,
    "interpretation": "TESVEN's frozen evaluation asserts each decision against this corpus oracle; this artifact preserves aggregate TESVEN counts only. Casewise agreement is established through the shared frozen per-case oracle, not by comparing stored TESVEN case IDs."
  },
  "metrics": {
    "false_verified_rate": {
      "numerator": 0,
      "denominator": 25,
      "value": 0
    },
    "critical_failure_recall": {
      "detected": 25,
      "abstained": 4,
      "denominator": 29
    },
    "oracle_coverage": {
      "numerator": 65,
      "denominator": 93
    },
    "p50_ms": 0.0025440000000003238,
    "p95_ms": 0.024190000000004375,
    "hosted_cost_per_run": null,
    "human_agreement": null
  },
  "limitations": [
    "Synthetic fixed fixtures only; not live Dovren, production usage, or customer data.",
    "Direct deterministic rule baseline; no Langfuse OSS service was run.",
    "Matching this suite shows no incremental case-outcome lift over the conventional checks on these fixtures; it does not measure provenance, tenant isolation, auditability, or operational burden.",
    "The TESVEN input result stores aggregate statuses rather than case IDs; casewise agreement is derived through both runners' per-case checks against the independently frozen oracle."
  ]
}