# HTML Artifact Collaboration Checklist 2026

**Version:** 1.9  
**Updated:** 20 October 2026  
**Author:** Raunaq Naidu  
**Publisher:** HTML Docs  
**Canonical guide:** https://www.html-docs.com/guides/html-artifact-collaboration-checklist

This open, vendor-neutral checklist provides 15 reproducible tests for deciding whether a live HTML artifact is merely shareable or ready for review, revision, automation, and responsible discovery. The checklist does not publish vendor rankings or first-party product scores.

## Method

1. **Use one representative artifact.** Include headings, a table, a link, an image with alternative text, and one browser interaction.
2. **Test three roles.** Repeat the workflow as publisher, signed-out reviewer, and API, CLI, or MCP client.
3. **Save evidence.** Record the tested URL, date, screenshot or response, and documentation link for every result.
4. **Score conservatively.** Award 2 for a reproducible pass, 1 for a documented workaround, and 0 when the capability is missing or unverified.

## Scoring

- **2 — Reproducible pass:** the behavior works as described and has saved evidence.
- **1 — Documented workaround:** the outcome is possible through an extra, documented step.
- **0 — Missing or unverified:** the capability is absent or the result cannot be reproduced.

Interpret the total score as follows:

- **26–30 — Collaboration-ready:** the workflow survives review and revision with little manual handoff.
- **17–25 — Shareable with gaps:** the artifact works, but feedback or automation still depends on side channels.
- **0–16 — Fragile handoff:** treat the link as a preview, not a durable collaboration record.

## Checklist

| ID | Category | Criterion | Pass signal | Evidence test | Points |
|---|---|---|---|---|---:|
| A1 | Access | Stable HTTPS URL | The artifact opens at one durable HTTPS URL after revisions. | Save the URL, publish a revision, and confirm the original URL still resolves. | 2 |
| A2 | Access | Explicit access mode | Public, unlisted, and restricted states are distinguishable before sharing. | Record the selected state and test it in a signed-out browser. | 2 |
| A3 | Access | Rendered browser view | A reviewer can use the artifact without downloading or running local files. | Open the link on a second device and test its primary interaction. | 2 |
| R1 | Review | Contextual feedback | Feedback can point to a section, element, or selected passage. | Leave a comment on one specific part and confirm its context survives reload. | 2 |
| R2 | Review | Resolvable discussion | Comments support replies and an explicit resolved or reopened state. | Reply, resolve the thread, reload, and reopen it. | 2 |
| R3 | Review | Version recovery | Prior versions are identifiable and can be inspected or restored. | Make two changes, identify both versions, and restore the first. | 2 |
| G1 | Agent continuity | Machine-readable state | A documented interface can read the artifact and its review state. | Save one successful API, CLI, or MCP response and its documentation URL. | 2 |
| G2 | Agent continuity | Scoped updates | Automation can update a defined region without republishing unrelated content. | Change one region and confirm an unrelated region is byte-for-byte unchanged. | 2 |
| G3 | Agent continuity | Durable identifiers | The document and editable regions keep stable identifiers across revisions. | Compare identifiers before and after an update and record any remapping. | 2 |
| G4 | Agent continuity | Concurrent update safety | Parallel edits merge safely or a stale write is rejected before it overwrites newer work. | Open two clients at the same revision, save independent edits, then attempt one stale write and confirm the system preserves both accepted changes or surfaces an explicit conflict. | 2 |
| P1 | Provenance | Authorship and revision provenance | Each tested revision records a stable identifier, generation time, and responsible human or software agent. | Publish one change, then save the revision identifier, timestamp, and attributable actor or agent from the UI or machine-readable history. | 2 |
| T1 | Trust and discovery | Indexing intent | Canonical URL and index or noindex intent are visible and consistent. | Inspect the canonical link and robots directive on the rendered page. | 2 |
| T2 | Trust and discovery | Share preview metadata | Title, description, image, and canonical page URL are declared for sharing. | Inspect the Open Graph properties and validate the image URL. | 2 |
| T3 | Trust and discovery | Accessible reading path | Headings, landmarks, keyboard use, contrast, and text alternatives are tested. | Run keyboard and automated checks, then record remaining manual findings. | 2 |
| T4 | Trust and discovery | Active content boundary | Scripts and embedded resources run inside a documented security boundary. | Record the content policy or isolation model and test one blocked resource. | 2 |

## Standards and source notes

- [Google Search: canonical URLs](https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls) — canonical selection and duplicate URL handling.
- [Google Search: robots meta tags](https://developers.google.com/search/docs/crawling-indexing/robots-meta-tag) — explicit index and noindex controls.
- [Open Graph protocol](https://ogp.me/) — minimum metadata for link previews.
- [W3C Web Content Accessibility Guidelines 2.2](https://www.w3.org/TR/WCAG22/) — accessibility success criteria and conformance.
- [W3C PROV-O](https://www.w3.org/TR/prov-o/) — interoperable attribution of entities and revisions to human or software agents.
- [IETF RFC 9110: Conditional Requests](https://www.rfc-editor.org/rfc/rfc9110.html#section-13) — preconditions and validators that prevent parallel clients from overwriting accepted changes.
- [OWASP Content Security Policy Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/Content_Security_Policy_Cheat_Sheet.html) — security boundaries for active HTML content.

## Reproduction files

- [CSV checklist](https://www.html-docs.com/research/html-artifact-collaboration-checklist-2026.csv)
- [JSON checklist](https://www.html-docs.com/research/html-artifact-collaboration-checklist-2026.json)
- [Markdown audit worksheet](https://www.html-docs.com/research/html-artifact-collaboration-audit-template-2026.md)
- [JSON audit schema](https://www.html-docs.com/research/html-artifact-collaboration-audit-schema-2026.json)
- [Valid blank audit example](https://www.html-docs.com/research/html-artifact-collaboration-audit-example-2026.json)
- [BibTeX](https://www.html-docs.com/research/html-artifact-collaboration-checklist-2026.bib)
- [Citation File Format](https://www.html-docs.com/research/html-artifact-collaboration-checklist-2026.cff)

Validate a completed audit with the pinned AJV CLI:

```bash
curl -fsSLO https://www.html-docs.com/research/html-artifact-collaboration-audit-schema-2026.json
npx --yes --package=ajv-cli@5.0.0 --package=ajv-formats@3.0.1 \
  ajv validate --spec=draft2020 -c ajv-formats \
  -s html-artifact-collaboration-audit-schema-2026.json \
  -d your-audit.json
```

## Citation

Naidu, Raunaq. “HTML Artifact Collaboration Checklist 2026.” HTML Docs, version 1.9, updated 20 Oct. 2026. https://www.html-docs.com/guides/html-artifact-collaboration-checklist

Markdown:

```markdown
[HTML Artifact Collaboration Checklist 2026](https://www.html-docs.com/guides/html-artifact-collaboration-checklist) — Raunaq Naidu, HTML Docs, version 1.9, updated 20 Oct. 2026.
```