Round offline-contract-audit
Offline contract audit
Source-reported narrativeRound README and results text are source material. Numeric values that are not tied to a claim ledger record remain source-reported.
Round source README · rounds/offline-contract-audit/README.md
Round date: HISTORICAL (before 2026-10-09) Publication badge: DESCRIPTIVE; KEPT FOR AUDIT Why not VALID: The historical inventory retains DESCRIPTIVE because of POST_HOC_ANALYSIS, RAW_EVIDENCE_MISSING. See the round evidence and limitations below. Recomputation status: OBSERVED SOURCE ONLY
Recomputation limit: Published summary figures: planned/observed counts, outcome totals/rates, and numeric estimates as present in README.md and RESULTS.md. These figures cannot be recomputed from repository inputs because no public per-round analysis script with a complete, identified input bundle is published. See the round README, MEASURED.md, and MISSING.md for recorded scope and gaps; historical execution inputs/toolchains/grading are not bundled.
Status
DESCRIPTIVE
Why not VALID
- POST_HOC_ANALYSIS: This is an offline blinded replay without a repair or Build rerun.
- RAW_EVIDENCE_MISSING: The contract-level raw records are absent.
Required reproduction metadata
- Kogen commit: not recorded in the cited round source.
- Harness commit: not recorded in the cited round source as a Git commit; any adapter or runtime fingerprints are not Git commits.
STATUS: INTERIM — Offline blinded replay; no repair or Build rerun. Study figures in the result section are source-reported, not reproducible from public data. The accounting table classifies this as an offline analysis; its zero live Build counts are not measured outcomes. The contract-level audit records are not included in this public page.
Scope and reported result
The audit compared Sol medium and Luna max on five known R48 false accepts plus passing-contract controls. The source report records 40 audits over 20 recoverable contracts. Sol medium identified the exact defect in 1 of 5 cases; Luna max identified 0 of 5. On passing contracts, the audit flagged 10 of 10 for Sol medium and 6 of 10 for Luna max. The report gives approximate per-audit costs of $0.049–$0.050 for Sol medium and $0.0042–$0.0044 for Luna max.
The controls measure a flagging/noise proxy. Because no repair or Build rerun was performed, the result does not show that contract review prevents false acceptance or improves live completion.
Live Build accounting
| planned_n | started_n | finished_n | graded_n | itt_n |
|---|---|---|---|---|
| 0 | 0 | 0 | 0 | 0 |
This page covers an offline audit of saved contracts, not a live Build cohort. Audit calls and contract cases are separate analysis units.
Reproduction record
| Field | Record |
|---|---|
| Harness / Kogen commit | No live harness or Kogen execution is described. Revisions for the historical R48 contracts are not listed in the supplied public input. |
| Model and effort | Sol medium and Luna max are source-reported; exact model builds are not supplied. |
| Task IDs | Exact R48 false-accept and control IDs are not supplied in a public crosswalk. |
| Command | The exact audit command is not recovered from the supplied public input. |
| Raw records | The source inventory names OFFLINE-CONTRACT-AUDIT.md and an offline-contract-audit record set; the contract-level raw records are not included in the public data tree. |
Until the contract IDs and audit records are published, all figures remain source-reported, not reproducible from public data.
Captured run records
Standard-schema records assigned to this round in the indexed snapshot.
No per-round file is listed. Check the run-record index for the complete partition list.
Reproduction inventory
Published analysis replay and full experiment replay are reported separately.
- Replay type
unavailable- Script
- No per-round analysis script is published.
- Inputs
- No identified public input bundle.
- Full execution replay
- Unavailable
- Scope
- No public per-round analysis script with a complete, identified input bundle is published. See the round README, MEASURED.md, and MISSING.md for recorded scope and gaps; historical execution inputs/toolchains/grading are not bundled.
View the complete per-round reproduction inventory · Repository reproduction guide