# AUD-001 evidence pack — V3 launch-gate record correction

## Work-package charter

- **Task IDs:** `AUD-001`
- **Owner:** Oshun portfolio audit-integrity maintainer; implementation owner:
  Codex
- **Accountable product:** V1–V9 autonomous-content portfolio, with V3 as the
  affected product
- **Priority:** P0
- **Target milestone:** 2026-07-15 SOTA gap-closure program
- **Risk tier:** High audit-integrity risk; no production runtime mutation
- **Dependencies:** `V3/launch/launch-readiness-release-gate.json`, the V3
  launch-readiness manifest, and the `@oshun/analytics` verification suite
- **Affected schemas and content slots:** `v3.launch-readiness-release-gate.v1`;
  no content slot or shipped artifact is changed
- **Rollout strategy:** Correct only documents that present the superseded count
  as the current state, then enforce the values from the committed JSON artifact
  in a value-asserting regression test.
- **Rollback strategy:** Revert the task-scoped commit. A rollback must not
  restore a prose value that disagrees with the committed evidence artifact.
- **Evidence-pack location:** This file

## Result

`AUD-001` is complete at implementation commit
`d223d3e65696e8bd1532bf6a0c5c12987f050263`.

- The canonical V1–V9 audit, the V3 autonomous-content audit, and the dated
  V1–V7 derived audit view report **30/41 required commands**, **24/39 named
  gates**, `red`, and `gaReady: false`.
- The regression test reads the committed machine artifact and derives the
  expected prose at test time. The counts are not duplicated in the assertion.
- A deliberately stale `29/40` fixture fails the same consistency predicate.
- Repository search found no generated HTML that presents `29/40` as current.
  The remaining occurrences are the adversarial review's quotation of the
  defect, this task ledger, a genuine intermediate historical measurement in
  `V3/REMEDIATION_2026-06-12.md`, and the negative fixture.

This result proves record consistency. It does **not** assert that the June V3
evidence is fresh or that V3 is GA-ready. The missing source commit and stale
`generatedAt` remain explicitly owned by `AUD-002`.

## Environment and immutable inputs

| Field              | Value                                                              |
| ------------------ | ------------------------------------------------------------------ |
| Repository commit  | `d223d3e65696e8bd1532bf6a0c5c12987f050263`                         |
| Host               | `dev-server`                                                       |
| OS/kernel          | Ubuntu Linux, `6.8.0-110-generic`, x86_64                          |
| Node.js            | `v24.14.1`                                                         |
| pnpm               | `10.25.0`                                                          |
| Vitest             | `4.1.7`                                                            |
| ESLint             | `10.4.0`                                                           |
| TypeScript         | `6.0.3`                                                            |
| Lockfile SHA-256   | `c0be8b29906dc32eff749a337329981e2ca258b69382eec835c677e10d44b4f4` |
| Model/provider use | None; this work package performs no inference or provider call     |

## Verification transcript

All timestamps are UTC on 2026-07-15. Each command ran from the repository root
at the implementation commit.

| Start        | End          | Exit | Command/result                                                                                                                                                                                                                                                                                                                              |
| ------------ | ------------ | ---: | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 12:08:08.581 | 12:08:10.696 |    0 | `pnpm exec vitest run --config libs/oshun/analytics/vitest.config.ts libs/oshun/analytics/src/__tests__/v3-launch-readiness.test.ts --maxWorkers=1` — 1 file, 7 tests passed                                                                                                                                                                |
| 12:08:10.698 | 12:08:15.609 |    0 | `pnpm exec eslint --no-ignore libs/oshun/analytics/src/__tests__/v3-launch-readiness.test.ts`                                                                                                                                                                                                                                               |
| 12:08:15.610 | 12:08:19.681 |    0 | `pnpm exec tsc --noEmit -p libs/oshun/analytics/tsconfig.lib.json`                                                                                                                                                                                                                                                                          |
| 12:08:19.681 | 12:09:27.887 |    0 | `pnpm nx build @oshun/analytics` — analytics and i18n build targets succeeded; final run used the local Nx cache                                                                                                                                                                                                                            |
| 12:09:38.629 | 12:09:40.761 |    0 | `pnpm exec prettier --check V1_V9_AUTONOMOUS_CONTENT_SOTA_AUDIT_2026-07-15.md V3_AUTONOMOUS_CONTENT_AGENTS_AUDIT_2026-07-14.md V1_V7_PLAN_SET_AUDIT_2026-06-12.md libs/oshun/analytics/src/__tests__/v3-launch-readiness.test.ts libs/oshun/analytics/src/__tests__/fixtures/v3-launch-readiness-stale-record.md evidence/v1-v9/AUD-001.md` |
| 12:09:40.761 | 12:09:41.112 |    0 | `git diff --check`                                                                                                                                                                                                                                                                                                                          |

The successful implementation commit hook also ran staged ESLint/Prettier, V2
asset locking, V5 documentation validation, the staged stub scan, and affected
typechecks for `@oshun/analytics`, `@oshun/mobile`, `@oshun/web`, and
`@oshun/bff`.

## Value, contract, and adversarial coverage

- **Value/integration:** The positive test opens the real committed JSON and all
  three current-state Markdown records, derives the expected statement from
  `releaseGate`, and requires every record to match.
- **Contract:** The pre-existing test continues to verify the v1 schema and
  manifest, derives both count families from their maps, and proves red/green
  and `gaReady` consistency.
- **Negative/adversarial:** `v3-launch-readiness-stale-record.md` contains the
  original `29/40` defect and is required not to match current evidence.
- **E2E/UI/accessibility/performance:** Not applicable. No runtime route,
  customer UI, accessibility behavior, or performance boundary changed.
- **Human review/signature:** No policy requires a human authority decision for
  correcting a machine-derived audit count. No reviewer identity or signature is
  fabricated.

## Artifact identities

| Artifact                                                                               | SHA-256                                                            |
| -------------------------------------------------------------------------------------- | ------------------------------------------------------------------ |
| `V3/launch/launch-readiness-release-gate.json` (`v3.launch-readiness-release-gate.v1`) | `3ae60abc2d84f0887e1125fbb9f7077e4b7d1e39eacb4ea35624cbebc661e754` |
| `V1_V9_AUTONOMOUS_CONTENT_SOTA_AUDIT_2026-07-15.md`                                    | `330f80b12d2547b81aff313e99cef04fa531def4f39a00899db5f57d9b358207` |
| `V3_AUTONOMOUS_CONTENT_AGENTS_AUDIT_2026-07-14.md`                                     | `3205900508b9fa889e7013ad5686baf07b3faeb7554752bf85dd9b9f067b2150` |
| `V1_V7_PLAN_SET_AUDIT_2026-06-12.md`                                                   | `7c5a2a36dd869add85c21c5db46c29b13da0be04d757f642990abdf70161633b` |
| `libs/oshun/analytics/src/__tests__/v3-launch-readiness.test.ts`                       | `171e6750c83831b9261d95c72c88ab89cead4e717731d50950c506c4b358f65d` |
| `libs/oshun/analytics/src/__tests__/fixtures/v3-launch-readiness-stale-record.md`      | `3ee2b95c1796977acefaaf0b0309f9b18e3863fbefd1d91deddade2fa7f30c88` |

No content artifact, source/canon/policy/rubric version, release placement,
screenshot, media probe, or runtime receipt changed in this documentation-only
work package.

## Broader repository-health observations

The non-acceptance discovery run
`pnpm exec vitest run --config libs/oshun/analytics/vitest.config.ts --maxWorkers=1`
completed with 73/79 files and 848/856 tests passing. Its eight failures are
pre-existing inventory drift outside `AUD-001`: three missing legacy Playwright
files referenced by V1 manifests, one renamed domain-switcher test title, and
one workflow display-name mismatch (with duplicate manifest consumers). The
focused launch-readiness file passed 7/7 in that run and in the timestamped
acceptance run above. The full spec `tsc` likewise reports existing errors in
unrelated negative-fixture casts and older test imports; production-library
typecheck and the affected commit-hook typechecks are green.

These observations are not used to turn `AUD-001` green and do not weaken its
acceptance check. They remain visible follow-up test-coherence work rather than
being silently omitted.
