# Local validation receipt

Date: 2026-10-25. Package D30-F04-A02. Local video-ready private draft; no remote publication. Raw process output is retained in local-validation.json (retained in the canonical package tests/evidence directory); actual browser outcomes in visual-qa.json (retained in the canonical package tests/evidence directory).

- 10 Python test methods pass, including 62 independent before-core financial/status expectations, input immutability, exact trace reconstruction, strict containers, all-closed chart history, nonsemantic array permutations and pinned historical validator prefix hashes.
- 62 complete Trial outputs and 84 complete unchanged historical-fold outputs agree between actual Python and TypeScript executions.
- Actual Playwright Chromium checks pass at 390, 768 and 1440 pixels: 13 scenarios per width, 39 runs, 270 authored receipt/report-state comparisons. The browser executes the embedded compiled TypeScript core and compares complete outputs against the standalone compiled core.
- 186 complete browser Trial outputs additionally pass the independent fixture expectations; 252 complete unchanged fold outputs match the reference across the three widths.
- Canonical September columns 125000, October cumulative columns 100000 versus movement 150000, historical p2/h4, exact atomic correction p3/p5, closed expense history, negative cash, huge integers, missing zero accounts and both financial failure kinds are explicitly asserted.
- Copied hypothetical reports change only report values and IDs; accepted source journals remain unchanged. Transport, speed, pause persistence, keyboard stepping, scenario/reset, terminal disabling and reduced motion pass. Every control target is at least 44 pixels. No horizontal overflow, console errors or network requests.
- Three native SVGs and Mermaid workflow render successfully. Actual inspection covers full SVG screenshots, 390-pixel top/full lab and the complete workflow. Text, amounts and account columns are legible without clipping; mobile audit cards wrap within the viewport.
- Full package validator: zero errors and zero warnings. Guided-playground validator: pass.

Figures: trial columns (retained in the canonical package tests/evidence directory), booking periods (retained in the canonical package tests/evidence directory), balanced wrong projection (retained in the canonical package tests/evidence directory), workflow (retained in the canonical package tests/evidence directory). Browser screenshots: 390-pixel top (retained in the canonical package tests/evidence directory), 390-pixel full (retained in the canonical package tests/evidence directory), 768-pixel full (retained in the canonical package tests/evidence directory), 1440-pixel full (retained in the canonical package tests/evidence directory).

The report vectors at each lab receipt boundary are synthetic independently authored teaching inputs, not outputs recycled as expected data. The independent 62-case oracle generator imports no algorithm core. Cross-language parity and browser/reference agreement supplement those financial expectations; neither alone establishes economic truth.

## Visitor integration

Root visitor integration pending. Package-local evidence does not claim a production-site or remote-publication result. The parent integration owner will append the isolated visitor-build receipt.