# Owned publication manifest

## Release target

- Route: `/writing/same-system-propose-approve-verify/`
- Canonical: `https://jaredmodstorage.github.io/writing/same-system-propose-approve-verify/`
- Publisher: Jared Mastroianni personal authority site only
- Publisher owner: Jared Mastroianni
- Schema: `TechArticle`, `ImageObject` and `BreadcrumbList`
- Author: Jared Mastroianni
- State: publication-ready local package only; no publisher handoff or public mutation
- Excluded owned publisher: modSTORAGE WordPress must reject the same complete article, title, slug, body, fictional Ashbury Gate example and practical tool as a duplicate

## Required public package

- `the-same-system-should-not-propose-approve-and-verify.md`
- `ai-decision-separation-register.csv`
- `source-register.csv`
- `054-ai-development-computer-vision-pilot-inspection.png`
- `image-notes.md`

## Presentation requirements

- Preserve the title, deck, byline, editorial disclosure and source limitations.
- Place the governed image after the byline with the exact alt text and caption from `image-notes.md`.
- Present the CSV as a downloadable operator tool and retain its blank template and six explicitly fictional rows.
- Preserve proposal, authorization, execution, provider receipt, governing readback and reconciliation as separate states.
- Preserve the approval digest, authenticated principal, role, effect, scope and expiration as one bound decision record.
- Do not describe the proposed separation layer as a Facily capability, deployed control, legal conclusion, compliance result, certification, audit, safety guarantee or measured operating result.

## Separate release gates

1. Reconfirm the title, route, slug, fictional Ashbury Gate identity and register are absent immediately before any publisher mutation.
2. Verify every governed hash against `checksums.sha256`.
3. Build and test the source locally; verify structured data, internal links, source links, download and image.
4. Commit source with a known rollback parent.
5. Deploy once through the established personal-authority-site workflow.
6. Read back the canonical, H1, byline, disclosure, image, download, metadata, sitemap and feed at desktop and mobile widths.
7. Record source, deployment, rollback and provider-run identities separately.
8. Keep publication, crawling, indexing, ranking, coverage and recognition as separate states.
