docs-navigation-linking — Navigation and Linking Analysis

Workflow#

1. Gather the docs — get the list of pages in scope and read their content. If a semantic/graph search tool over the markdown is available (self-hosted markdown-lsp, or a connected Docsbook workspace), prefer it — faster and cheaper than scanning files; otherwise read the files directly with grep/find. Prioritize Tier 1 pages (quick-start, pricing, auth, install) first. This skill requires the full doc graph — do not run on a single page in isolation.
2. Build link sets — extract all href values from every page; compare to the full page list to identify orphans and broken links.
3. Apply checklist — check internal links, orphan pages, anchor text quality, navigation hierarchy, next steps / cross-references, and external links.
4. Produce report — return one JSON issue object per finding, sorted by severity.

Guardrails#

• Do not edit any documentation files — surface findings only.
• This skill is cross-file: orphan detection requires the full graph. Never run on a single page in isolation.
• The root index/home page is expected to have no inbound links from docs — exclude it from orphan detection.
• Checking external link validity requires HTTP requests — ask the user before making outbound calls.
• Confirm Tier 1 pages with the user before flagging navigation accessibility as medium vs. high severity.

Inputs#

This skill needs two things, by whatever means are available:

• The list of pages in scope — a docs folder, a sitemap, or a doc graph.
• The content of each page — read on demand.

Acceleration (optional). Graph/semantic search over the docs makes navigation faster and cheaper than scanning files. You can self-host it with markdown-lsp, or get the same capability in the cloud by connecting a Docsbook workspace. With nothing connected, plain file reads and grep/find work fine.

Checklist#

Internal Links#

• All referenced pages exist in the doc graph — no broken internal links
• Relative paths are consistent — avoid mixing /docs/ absolute paths with ../ relative paths
• Anchor links (#section) resolve — the heading they target exists on the page
• Linked pages are relevant — links add value, not just bulk

Orphan Pages#

• Every page has at least one inbound link from another doc page
• Exception: root index/home page (expected to have no inbound from docs)
• Orphan detection: pull all href values from every page, compare to full page list

Anchor Text Quality#

• No "click here", "read more", "here", "this link", "this page"
• Anchor text describes the destination — "see the authentication guide" not "see here"
• No full URLs as anchor text — https://example.com/long-path → use descriptive text
• Consistent naming — if the target page is called "Custom domain setup", call it that in all links to it

Navigation Hierarchy#

• Maximum 3 levels deep for most content (sidebar: section → subsection → page)
• Tier 1 pages are accessible in 1 click from homepage or top navigation
• Related pages are grouped — all guides together, all reference together
• No dead-end pages — every page has a "next step" or at minimum returns to a parent

Next Steps and Cross-References#

• Tutorials end with "Next steps" linking to related how-to guides
• How-to guides link to related reference pages for the commands they use
• Reference pages link back to the how-to or tutorial that demonstrates them
• Concept pages link to the tutorial that puts the concept into practice

External Links#

• External links are stable — prefer official documentation, not blog posts that may disappear
• No broken external links — check periodically; flag pages with external links in the report
• External links open in a new tab (if controlled at markdown level) or are explicitly labeled

Orphan Detection Pattern#

Using the doc graph (the page tree + links):

1. Build a set of all page paths: all_pages = {page.path for page in graph}
2. Build a set of linked pages: linked = {href for page in graph for href in page.links}
3. Orphans: all_pages - linked - {homepage}

What to Look For#

Output Format#

{
  "file": "docs/guides/advanced-auth.md",
  "line": null,
  "severity": "high",
  "rule": "orphan-page",
  "found": "docs/guides/advanced-auth.md has no inbound links from any other documentation page. It is unreachable through navigation — only discoverable via direct URL or search.",
  "suggestion": "Add a link from docs/guides/authentication.md: 'For token rotation and HMAC signatures, see [Advanced authentication](/docs-skills/skills/advanced-auth)'. Also add to the sidebar navigation."
}

{
  "file": "docs/quick-start.md",
  "line": 67,
  "severity": "critical",
  "rule": "broken-internal-link",
  "found": "Line 67: [Connect your GitHub repo](/docs-skills/skills/setup/github-integration) — the target file docs/setup/github-integration.md does not exist in the doc graph.",
  "suggestion": "Either: (1) Restore the missing page, or (2) Update the link to the correct path. Check git history for a rename or deletion."
}

{
  "file": "docs/api/workspaces.md",
  "line": 23,
  "severity": "high",
  "rule": "poor-anchor-text",
  "found": "Line 23: 'For more information, [click here](/docs-skills/skills/guides/workspaces).' — 'click here' provides no context about the destination.",
  "suggestion": "Replace with descriptive anchor: 'See the [workspace configuration guide](/docs-skills/skills/guides/workspaces) for full details.'"
}

{
  "file": "docs/tutorials/first-deploy.md",
  "line": null,
  "severity": "medium",
  "rule": "missing-next-steps",
  "found": "Tutorial 'Deploy your first site' ends abruptly with no 'Next steps' section. Users who complete it have no guided path forward.",
  "suggestion": "Add a 'Next steps' section at the end: '## Next steps\\n- [Set up a custom domain](/docs-skills/skills/guides/custom-domain)\\n- [Enable AI chat](/docs-skills/skills/guides/ai-chat)\\n- [View your analytics](/docs-skills/skills/analytics/overview)'"
}

Acceptance Criteria#

• The full doc graph has been used — orphan detection covers all pages, not a sample.
• All broken internal links include the exact file path and target that does not exist.
• External link checks are either completed or explicitly noted as skipped (no outbound calls without user confirmation).
• Output is valid JSON per the format above, one object per finding.

Related Skills#

• docs-seo — orphan pages have no inbound links = SEO problem too
• docs-accessibility — anchor text quality is an a11y requirement
• docs-maintenance — broken links are a maintenance issue
• docs-analyze — orchestrator