# Validation receipt — D30-F03-A05

Date: 2026-10-25. Private local package; no remote publication.

## Executable acceptance

- Python: 11 methods passed, including 65 pre-implementation literal financial/status cases.
- Python/TypeScript: all 65 complete proof outputs compare exactly.
- Pinned common core: all 105 complete independent selection outputs unchanged.
- Additional checks: 512-version accepted chain and 513-version guard, canonical negative/huge controls, chosen EUR with external-outflow adjustment sign, Unicode scalar/whitespace reasons, input/output detachment, future control payload hiding, and seeded delivery/register permutations.
- Both full-package and self-contained-playground structural validators pass.
- Raw command/exit/stdout/stderr evidence: [acceptance-results.json](evidence/acceptance-results.json).

The proof authoring script imported no runtime implementation. Its 65 fixed expected financial facts, outcomes and selected-control assertions were materialized before the core. Full-output language parity supplements those independent assertions; it is not described as independently generated complete proof JSON. The shared selection dataset separately has complete independent expected outputs.

## Actual browser and visual QA

Local Playwright Chromium ran 13 guided scenarios at 390, 768 and 1440 pixels: 39 scenario runs and 273 complete receipt-stage comparisons against the actual reference kernel. In each width, all 65 complete proof outputs were compared and their literal independent assertions checked: 195 full browser/reference proof comparisons. All 105 complete independent shared selection expectations ran in each width: 315 further checks.

Passed controls: corrected explained-but-unresolved preview; earlier external closing 14400 versus corrected 15000 at equality; missing source versus known internal failure; register omission; copied internal close +1; null opposite residual; all 15 supplied source/control versions retained in full audit; Begin/Back/Step/Play/Pause/Speed/Reset; keyboard arrows; reduced-motion single-step playback; minimum 44-pixel controls; no horizontal overflow; no console errors or network requests.

Raw browser receipt: [visual-qa.json](evidence/visual-qa.json). Captures: [390 top](evidence/lab-390-top.png), [390 full](evidence/lab-390.png), [768 full](evidence/lab-768.png), [1440 full](evidence/lab-1440.png).

Two native SVGs and the actual Mermaid workflow were rendered and visually inspected. Text, signs, zero residuals and indeterminate status remain legible without clipped or overlapping labels. Captures: [two bridges](evidence/two-source-bridge.png), [residual/status](evidence/residual-and-status.png), [workflow](evidence/balance-proof-flow.png).

## Trust and release boundary

A clean result means arithmetic consistency under supplied independently trusted heads and controls. The deliberately forged-host-claim case passes arithmetic and is explicitly labeled outside authentication. No source identity, real-world completeness, finality, investigation authority, journal posting or suspense clearing is established by this package.

The canonical package remains video-ready/private; the audiovisual deliverable is a scene script. Root visitor integration receipt is pending and will be appended after isolated website preview acceptance. No shared website or release manifest was edited by this topic owner.