# Bible Timeline V23.1.26 — Twenty-Fifth-Pass Engineering Audit

## Scope

Defect correction and release stabilization only. No new historical or theological claims were introduced during this pass.

## Corrected defect classes

1. **Teaching/Governance fallback cross-tab data loss.** Fallback mutations refresh and merge the latest persisted aggregate before writing, preventing stale tabs from erasing unrelated records.
2. **Advanced Atlas cross-tab workspace data loss.** Saves, deletes, loads, exports, imports, and storage-event refreshes now act on the current persisted workspace library.
3. **Generated lesson duration violations.** Lesson targets normalize to 10–180 minutes, generated event blocks to 1–30 minutes, and malformed saved targets to 45 minutes.
4. **Stale active-release metadata.** Current documentation, release status, Crucifixion Chronology branding, and audit links identify V23.1.26.
5. **Incomplete retained regression registration.** Historical browser contracts through the twenty-fourth pass remain registered in the archival retained command.
6. **Oversized retained browser lifetimes.** Fourth- and sixth-pass harnesses use bounded browser groups rather than one long-lived monolithic Chromium session.
7. **Post-biblical taxonomy leakage.** Raw hyphenated era identifiers were replaced with readable labels and regression coverage.
8. **Post-biblical page separation.** Eighty-eight post-biblical events now occupy their dedicated timeline without inflating the biblical master chronology.
9. **Production-only interface residue.** The lawful-access blocker artifact was removed from the release.
10. **Primeval chronology placement.** Primeval research was removed from the dated event sequence and retained on its dedicated page, leaving 148 visible master-timeline events.
11. **Unsupported Early Exodus display mode.** The Early Exodus model was removed from the main chronology selector; its Error Dossier and critical research remain intact.
12. **Initial-load overfetching.** The large V10 expansion and artifact datasets now load on demand when a dossier or artifact feature needs them.
13. **Package-wide service-worker precaching.** Installation now verifies a 34-file application shell and caches specialist content at runtime, preserving rollback protection without redownloading the full release on each revision.
14. **Stale release validators.** Event-count, cache-architecture, specialist-offline, and release-metadata assertions now match the active implementation.
15. **Incomplete generated registries after homepage script reduction.** Integrity and entity builders use explicit canonical dataset lists rather than inferring the project corpus from homepage script tags.
16. **Knowledge-graph health mismatch.** Active release metadata now agrees with the reconstructed graph at 333 nodes and 777 edges, eliminating the false Data Health failure.
17. **Obsolete retained Early Exodus interactions.** Historical contracts no longer attempt to select the removed Early Exodus main-timeline option; the dossier remains covered separately.

## Permanent regression coverage

- `scripts/v23-1-26-performance-budget-smoke.mjs`
- `scripts/v23-1-26-lazy-timeline-browser-smoke.py`
- `scripts/v23-1-26-data-health-browser-smoke.py`
- `scripts/v23-1-26-post-biblical-label-smoke.mjs`
- `scripts/service-worker-v23-1-26-smoke.mjs`
- `scripts/v23-1-26-cross-tab-store-smoke.mjs`
- `scripts/v23-1-26-atlas-cross-tab-browser-smoke.py`
- `scripts/v23-1-26-lesson-bounds-smoke.mjs`
- Hardened `scripts/v23-1-5-fourth-pass-browser-smoke.py`
- Hardened `scripts/v23-1-7-sixth-pass-browser-smoke.py`
- `scripts/v23-1-26-imperial-control-rail-smoke.py`
- Updated specialist offline contract in `scripts/astronomy-workbench-browser-smoke.py`
- `scripts/validate-v23-1-26.mjs`

## Final machine-backed result

The stabilized package contains 70 active pages, 41 registered programs, 56 exports, 86 JavaScript modules, and a 34-file required offline shell. It exposes 148 visible biblical-timeline events, 88 post-biblical events, and three selectable main chronology models while retaining four analytical model records. The canonical outputs contain 2,066 entity profiles, 7,393 connections, 569 passage contexts, 77 guided-research steps, and a 333-node / 777-edge knowledge graph.

All 10 publication outputs matched isolated rebuilds, and all 44 active generated outputs rebuilt byte-identically. Data Health reports 124 passes, 5 warnings, and 0 failures. Current semantic validation, the master-timeline browser gate, the specialist Birth/Astronomy/Consistency suites, the loading budget, lazy dossier hydration, service-worker rollback/runtime-cache behavior, cross-tab persistence contracts, responsive-layout contracts, the imperial control rail, retained compatibility checks, and editorial lint passed.

## Execution-window qualification

The audit environment can terminate a single monolithic regression command before all sequential browser processes finish. Every constituent current release-blocking command passed independently from the active tree. The 36 historical pass-two-through-pass-twenty-four browser contracts remain registered as a non-blocking archival suite; this report does not represent an uninterrupted rerun of that complete archive.