docs-analyzelisted
Install: claude install-skill Docsbook-io/docs-skills
# docs-analyze — Documentation Analysis Orchestrator
## Workflow
1. **Connect to Docsbook** — run `list_workspaces`, fetch the doc graph via `get_doc_graph({ format: "toon" })`. Offer `create_workspace` if repo is not indexed. Reindex with `reindex_doc_graph` if graph is empty or stale (> 7 days).
2. **Identify Tier 1 pages** — flag quick-start, pricing, authentication, and installation pages for priority analysis.
3. **Run sub-skills in parallel** — spawn independent Agent calls for `docs-content-types`, `docs-structure-templates`, `docs-style-tone`, `docs-audience`, `docs-seo`, `docs-accessibility`, `docs-maintenance`. Run `docs-navigation-linking` and `docs-i18n` sequentially (they depend on the full graph and workspace language settings).
4. **Aggregate and deduplicate** — collect JSON issues from all skills; merge cross-cutting findings (e.g. missing alt = a11y + SEO — report once under higher severity, note both skills).
5. **Produce final report** — output a prioritized markdown report with severity summary, critical issues, recommendations by area, and quick-win list.
## Guardrails
- Do not edit any documentation files — surface findings only.
- Do not run `docs-i18n` if only one language is enabled in workspace settings.
- A cross-skill finding (same line flagged by two skills) is reported once under the higher severity.
- Ask the user to confirm Tier 1 pages before starting — defaults may not match the project.
- If the graph is stale, offer to reindex before pr