{
  "id": "sample-audit-current@7",
  "schema": "woc.record.v1",
  "subject": "Word of Clout self-audit procedure — complete signed extract inventory for both published runs; verdicts carried unchanged from the predecessor record in the chain",
  "record_class": "self_audit",
  "material_class": "self_audit",
  "public_key": "-----BEGIN PUBLIC KEY-----\nMCowBQYDK2VwAyEAZAXc8TrIl05HTq4oDHbsc2Qf8SA6pydv3jbaUV4kSyA=\n-----END PUBLIC KEY-----",
  "propositions": [
    {
      "id": "self-audit:receipt_observable",
      "proposition": "The published receipt (sample-audit-n50-20260907T043637Z.out, sha256 4a763180b1aa2c92dc48cdd49f8016d0ac079846f8823e8f869aa5970497e83b) records 50 distinct verification ids: 42 board and 8 NPI. Under PROCEDURE-FAULT-1 v2 re-parse, 49 rows have expected == actual (agree); 1 row (row 13) carries a not-rederivable marker. Retained as published ledger extract at /verify/self-audit/2026-09-06T200705434Z/ledger/ (20/20 SHA256SUMS-verified) and /verify/self-audit/2026-09-07T043637637Z/ledger/ (50/50 SHA256SUMS-verified). Runner published at /verify/self-audit/observation-extract-runner-v2.mjs re-derives 49/50 agree + 1 NOT-REDERIVED (ver:board:mo-dental:2025005314:address_of_record:2026-09-04@old, PROCEDURE-FAULT-1 v2 rule-short-circuit).",
      "evidence_class": "receipt_observable",
      "state": "VERIFIED",
      "authority": "operator run receipt",
      "verified_at": "2026-09-22T04:27:00.000Z",
      "method": "recount of the published receipt; ledger extract published at /verify/self-audit/2026-09-07T043637637Z/ledger/ with SHA256SUMS; re-derivation runner at /verify/self-audit/observation-extract-runner-v2.mjs",
      "next_required_by": "2026-12-06T04:36:37.637Z",
      "decay_class": "unmeasured",
      "provenance": {
        "verified_at_source": "OBSERVED",
        "next_required_by_source": "POLICY"
      }
    },
    {
      "id": "self-audit:frame_unverifiable",
      "proposition": "The frame is every verification row recorded in the Word Of Clout published-records store at cutoff 2026-09-07T04:36:37.637Z. The frame row count is not recorded in the run receipt; the frame hash at the cutoff was not captured.",
      "evidence_class": "run_design",
      "state": "UNVERIFIABLE",
      "authority": "narrowed in the record chain: the published receipt cannot establish this property (missing precondition stated on the proposition)",
      "method": "UNVERIFIABLE - missing precondition: frame count/hash at cutoff not recorded",
      "missing_precondition": "frame count/hash at cutoff not recorded",
      "next_admissible_step": "a successor run that records the frame row count and frame hash at the cutoff",
      "provenance": {
        "next_required_by_source": "N/A"
      }
    },
    {
      "id": "self-audit:draw_unverifiable",
      "proposition": "The draw used MongoDB's server-side $sample aggregate (sample_audit.js:91) against the frame. The sampler source that ran (sampler 51b81d94, sample_audit.js sha256 51fdb0dd3568cbe8faba3798a86428d341f66b052221fc9daf0075416333827d) is published at /verify/self-audit/sample_audit.js so the method can be inspected; no seed was recorded in the receipt, so the draw cannot be reproduced.",
      "evidence_class": "run_design",
      "state": "UNVERIFIABLE",
      "authority": "narrowed in the record chain: the published receipt cannot establish this property (missing precondition stated on the proposition)",
      "method": "UNVERIFIABLE - missing precondition: no seed recorded; the draw cannot be reproduced",
      "missing_precondition": "no seed recorded; the draw cannot be reproduced",
      "next_admissible_step": "a successor run that records the seed, the RNG, and the draw algorithm in the receipt",
      "provenance": {
        "next_required_by_source": "N/A"
      }
    },
    {
      "id": "self-audit:fault_classification",
      "proposition": "Row 13 (ver:board:mo-dental:2025005314:address_of_record:2026-09-04@old) fault class NOT-REDERIVED / sub-cause rule-short-circuit - this is a RETROSPECTIVE classification assigned from the receipt marker and the sampler source, not an emission of the run itself.",
      "evidence_class": "retrospective",
      "state": "VERIFIED",
      "authority": "@4 narrowing - the classification is a later labeling, not a run emission",
      "verified_at": "2026-09-22T04:27:00.000Z",
      "method": "retrospective classification under PROCEDURE-FAULT-1 v2",
      "next_required_by": "2026-12-06T04:36:37.637Z",
      "decay_class": "unmeasured",
      "provenance": {
        "verified_at_source": "OBSERVED",
        "next_required_by_source": "POLICY"
      }
    },
    {
      "id": "self-audit:bound_conditional",
      "proposition": "0 disagreements in 49 re-derived rows; the exact one-sided 95% binomial upper bound on the frame error rate among re-derivable verification rows is 1 - 0.05^(1/49) = 5.93% approx 5.9%, conditional on the run design above (the 50 rows are the ones in the receipt; the frame, the draw, and the re-derivation for the fault row cannot be independently established for this run). The retained re-derivable artifact is the ledger extract at /verify/self-audit/2026-09-07T043637637Z/ledger/ with SHA256SUMS and the observation-extract-runner-v2.mjs published beside it.",
      "evidence_class": "receipt_observable",
      "state": "VERIFIED",
      "authority": "@4 narrowing",
      "verified_at": "2026-09-22T04:27:00.000Z",
      "method": "exact binomial one-sided 95% upper bound; ledger extract and runner at /verify/self-audit/ published as retained artifact",
      "next_required_by": "2026-12-06T04:36:37.637Z",
      "decay_class": "unmeasured",
      "provenance": {
        "verified_at_source": "OBSERVED",
        "next_required_by_source": "POLICY"
      }
    },
    {
      "id": "self-audit:run_inventory",
      "proposition": "2 published receipts: 951d0480bd92f162bf21986134e86f48ac3dbe47b37934b9a486042d50a0b776 published at /verify/self-audit/sample-audit-20260906T200704Z.out; 4a763180b1aa2c92dc48cdd49f8016d0ac079846f8823e8f869aa5970497e83b published at /verify/self-audit/sample-audit-n50-20260907T043637Z.out.",
      "evidence_class": "receipt_observable",
      "state": "VERIFIED",
      "authority": "@4 narrowing",
      "verified_at": "2026-09-22T04:27:00.000Z",
      "method": "published receipt catalogue",
      "next_required_by": "2026-12-06T04:36:37.637Z",
      "decay_class": "unmeasured",
      "provenance": {
        "verified_at_source": "OBSERVED",
        "next_required_by_source": "POLICY"
      }
    },
    {
      "id": "self-audit:run_completeness",
      "proposition": "The published run-inventory records 2 runs; run completeness rests on WoC host records — whether additional runs occurred is not independently verifiable from the published receipts alone.",
      "evidence_class": "run_design",
      "state": "UNVERIFIABLE",
      "authority": "narrowed in the record chain: the published receipt cannot establish this property (missing precondition stated on the proposition)",
      "method": "UNVERIFIABLE - run completeness rests on WoC host records",
      "missing_precondition": "run completeness rests on WoC host records",
      "next_admissible_step": "every run appends its receipt hash to the signed host snapshot at run time",
      "provenance": {
        "next_required_by_source": "N/A"
      }
    }
  ],
  "dependency_manifest": {
    "ruleset_version": "woc.records.ruleset.self-audit.v1",
    "evaluation_cutoff": "2026-09-22T04:27:00.000Z",
    "receipt": {
      "name": "sample-audit-n50-20260907T043637Z.out",
      "sha256": "4a763180b1aa2c92dc48cdd49f8016d0ac079846f8823e8f869aa5970497e83b",
      "bytes": 14731,
      "citation": "/verify/self-audit/sample-audit-n50-20260907T043637Z.out"
    },
    "extract_inventory": {
      "inventory_version": "woc.extract-inventory.v1",
      "note": "Complete file inventory of both published run ledger extracts: every file's name, sha256, and byte size (row counts for .jsonl files), each run's expected re-derivation outcome counts computed from the pinned bytes, the re-derivation reference implementation's own hash, and the omission-closed verification runner's hash. The signed record is the only trust root: a consumer verifies a downloaded runner against runner_v2.sha256 here before running it, and the runner refuses any missing file, extra file, changed byte, missing row, edited checksum list, modified re-derivation implementation, stripped inventory, or total mismatch — the count-only gate that accepted a deleted verification row is gone.",
      "verify_root": "/verify/self-audit/",
      "runner_v2": {
        "path": "/verify/self-audit/observation-extract-runner-v2.mjs",
        "sha256": "374d83ff139cc1bca654444110cf3c4827f4b7d3765539be8113e429f2371e93",
        "bytes": 31733,
        "note": "the omission-closed consumer runner; a downloaded copy is checked against this signed sha256 before it is run"
      },
      "rederive_runner": {
        "sha256": "d6961143fcc1c5fbd3dfc1d655aaa5981607196784cdd7a82b2f5e0c128fa533",
        "bytes": 8195,
        "note": "the re-derivation reference implementation published identically inside both run ledger directories; bound here at its own inventory entry so a modified copy cannot stand in for it (no self-attestation)"
      },
      "runs": [
        {
          "run_id": "2026-09-06T200705434Z",
          "ledger_path": "2026-09-06T200705434Z/ledger",
          "public_url": "/verify/self-audit/2026-09-06T200705434Z/ledger/",
          "files": [
            {
              "name": "README",
              "sha256": "64d89c25ca0fcb33bda0442f7f31c00746f70eee624056cc340dd114ab913183",
              "bytes": 102
            },
            {
              "name": "SHA256SUMS",
              "sha256": "0b8c3b5c76d62a7fed818048c01f4a4a0a848d18130dc3fa5953b543e2c47fa5",
              "bytes": 335
            },
            {
              "name": "observations.jsonl",
              "sha256": "3bbdb00c515562f451ad8346ad49919242104ea6ea70a791f44f15e821b4ed92",
              "bytes": 5806,
              "rows": 20
            },
            {
              "name": "propositions.jsonl",
              "sha256": "48790e15c12d9ff8569646c893bb19c004b896ef097cb0e9ca5d1d951971bc78",
              "bytes": 4630,
              "rows": 20
            },
            {
              "name": "rederive.mjs",
              "sha256": "d6961143fcc1c5fbd3dfc1d655aaa5981607196784cdd7a82b2f5e0c128fa533",
              "bytes": 8195
            },
            {
              "name": "verifications.jsonl",
              "sha256": "9cb70c59b5a179baa695195a2d717277c79a4af1ba5fa956eb67d96bd31acbf1",
              "bytes": 8851,
              "rows": 20
            }
          ],
          "expected_outcomes": {
            "agree": 20,
            "disagree": 0,
            "not_rederived": 0
          }
        },
        {
          "run_id": "2026-09-07T043637637Z",
          "ledger_path": "2026-09-07T043637637Z/ledger",
          "public_url": "/verify/self-audit/2026-09-07T043637637Z/ledger/",
          "files": [
            {
              "name": "README",
              "sha256": "64d89c25ca0fcb33bda0442f7f31c00746f70eee624056cc340dd114ab913183",
              "bytes": 102
            },
            {
              "name": "SHA256SUMS",
              "sha256": "aa09177a52e2e55e5125a6f0321916e5351c38484ef3bb04a8ee1b87a453f33c",
              "bytes": 335
            },
            {
              "name": "observations.jsonl",
              "sha256": "654d097772b84362ca9483374b13c68254525aa88f68da1e60520e120d57e881",
              "bytes": 14368,
              "rows": 50
            },
            {
              "name": "propositions.jsonl",
              "sha256": "772db988570f3e320b62e5e08182b4f6595646b267a3f4b3924c8645e6817e7c",
              "bytes": 11533,
              "rows": 50
            },
            {
              "name": "rederive.mjs",
              "sha256": "d6961143fcc1c5fbd3dfc1d655aaa5981607196784cdd7a82b2f5e0c128fa533",
              "bytes": 8195
            },
            {
              "name": "verifications.jsonl",
              "sha256": "72fa9fee88c17fc274b818c94fbf3efddcc4496389cc12a4ca40de3c1a6e9b67",
              "bytes": 23312,
              "rows": 50
            }
          ],
          "expected_outcomes": {
            "agree": 49,
            "disagree": 0,
            "not_rederived": 1
          }
        }
      ]
    }
  },
  "dependency_manifest_sha256": "76e5a5660ade3e7650232a87e3e15cbca109703363b88c8105ded2bb015de138",
  "signed_scope_note": "The signature covers exactly: id, schema, subject, created_at (the EVIDENCE CUTOFF — the newest evidence timestamp in the derived snapshot, NOT the issuance time), issued_at (the issuance time of THIS record, distinct from the evidence cutoff — a record can never acquire an earlier issuance time because its evidence is older), public_key, propositions (four-state verdicts only, POLICY next_required_by = min(90d interval, authority-stated expiry at 00:00 America/Chicago)), candidate_relations (evidence joins carrying no verdict), dependency_manifest, dependency_manifest_sha256, signed_scope_note, supersedes — every field except signature, signing, and derivation. The derivation block (the queries that produced this record) is display metadata, excluded from signing. dependency_manifest_sha256 is the SHA-256 hex of the canonical form of dependency_manifest (key-sorted, no whitespace) — the same recipe as this record's own canonical form (equal manifest values hash equally regardless of stored key order). The field names in this note are the signed field names — machine-replayable values, labeled as such; the sentences around them state their meaning.",
  "issued_at": "2026-09-28T05:00:29.351Z",
  "supersedes": [
    "sample-audit-current@6"
  ],
  "signature": "IsrCfpSOaMit3CwuyM0ixI1aFPt7wH0IwBRFA1AqmwWD6fUGGfcQ+5IZttRGMGEKrD5NAO9E8E+bsml/Szw6Bg==",
  "signing": {
    "algorithm": "Ed25519",
    "hash": "SHA-256",
    "canonical_form": "canonical form (key-sorted, no whitespace; signature, signing, and derivation excluded)",
    "signed_at": "2026-09-28T05:00:29.653Z",
    "note": "Signed with the production signing key; the public key below is derived from it (SubjectPublicKeyInfo format, byte-identical under DER encoding) and matches the authority key embedded in the active crosswalk generation. The canonical-form recipe is published at /laws/canonicalization-v1.md."
  }
}