prompts/docs/skills/zensical-docs/SKILL.md

name, description, argument-hint

name	description	argument-hint
zensical-docs	Plan, write, and improve high-quality documentation with Zensical. Use for information architecture, progressive discoverability, writing standards, navigation/search tuning, and feature-driven docs site configuration.	What are you documenting, who is the audience, and what Zensical features are in scope?

Zensical Documentation Authoring

Create documentation that is easy to discover, easy to scan, and easy to trust, then configure Zensical so the site reinforces those outcomes.

When to Use

• You are creating or restructuring docs in a Zensical project.
• You need a repeatable workflow for docs quality and discoverability.
• You want guidance that combines writing best practices with Zensical feature choices.
• You need a checklist-driven review before publishing.

Progressive Loading References

Load only the references required for the current task:

• Source map for official docs and APIs: ./references/index.md
• Zensical features and configuration links: ./references/zensical-features.md
• Theme customization, colors, icons, and extension patterns: ./references/theme-customization-and-icons.md
• Docs writing quality and style guidance: ./references/documentation-quality.md
• Information architecture and discoverability patterns: ./references/discoverability-and-ia.md
• Code-heavy API docs with mkdocstrings: ./references/code-heavy-docs-and-mkdocstrings.md

Project Bootstrap Rule

Always start a new documentation project with uv run zensical new.

• This command provides the baseline scaffolding you need for structure, configuration, and theme setup.
• Do not hand-build a new project skeleton when this command is available.

Related Skill Discovery

When the task is not just writing docs but creating or wiring a new MCP skill in this repository, use catalog discovery to load the bootstrap skill before drafting implementation steps.

1. Search the catalog with terms such as new skill, skill bootstrap, or scaffold skill.
2. Fetch the new-skill document through the catalog tool path.
3. Use that skill for runtime package, metadata, and mount wiring, then return here for documentation architecture and Zensical-specific authoring guidance.

This keeps implementation guidance and documentation guidance separated while still making both discoverable from one request.

Inputs To Collect

Collect these before writing. If missing, make explicit assumptions.

1. Audience: beginner, intermediate, advanced, or mixed.
2. User goals: top tasks users come to complete.
3. Product scope: features, versions, and deployment context.
4. Content constraints: release deadlines, localization, legal/compliance needs.
5. Navigation constraints: explicit nav vs implicit structure.
6. Site-level expectations: search quality, code example depth, diagrams, API docs.

Procedure

Step 0: Audit Current Documentation

Build a quick baseline of what exists now.

1. List docs sections and current nav structure.
2. Identify orphan pages (not in nav, weak internal linking, or no inbound links).
3. Identify stale sections (version drift, missing prerequisites, broken commands).
4. Capture recurring support questions and map each to a missing or weak doc page.

Completion check: you can name the top 5 current documentation gaps.

Step 1: Define Outcomes and Page Taxonomy

Organize content by user intent, not by internal team boundaries.

1. Split content into at least four buckets:
   • Learn (concepts and mental models)
   • Do (task/how-to guides)
   • Reference (API, config, CLI, schemas)
   • Troubleshoot (symptoms, diagnostics, fixes)
2. For each bucket, define success criteria and expected time-to-answer.
3. Add clear page entry criteria (what belongs here, what does not).

Completion check: every planned page maps to one primary user intent.

Step 2: Design Progressive Discoverability

Make docs progressively discoverable from overview to detail.

1. Start each section with an index/overview page that answers:
   • who this section is for
   • what problems it solves
   • where to go next
2. Keep page openings front-loaded:
   • first paragraph states purpose and outcome
   • first heading after intro is usually prerequisites or quickstart path
3. Add consistent wayfinding on every page:
   • links to prerequisite concepts
   • links to next steps
   • links to relevant reference sections
4. Prefer short sections, descriptive headings, and stable anchor names.

Completion check: users can navigate from high-level overview to exact procedure in 3 clicks or fewer.

Step 3: Author High-Trust Content

Use writing patterns that reduce ambiguity and failure.

1. Use imperative, task-oriented titles (for example: "Configure SSO with OIDC").
2. State assumptions and prerequisites before commands.
3. Provide copy-paste-safe examples and expected outputs.
4. Include rollback or recovery steps for risky operations.
5. Separate normative guidance (must/should) from optional patterns.
6. Call out version-specific behavior explicitly.

Completion check: each task page includes prerequisites, steps, expected result, and failure recovery.

Step 4: Apply Zensical Features Intentionally

Load ./references/zensical-features.md and choose features based on content needs. Load ./references/code-heavy-docs-and-mkdocstrings.md when the docs include Python APIs or other generated references.

Decision matrix:

• If docs are deep and section-heavy: enable navigation.indexes, navigation.path, and navigation.sections.
• If speed and perceived responsiveness matter: enable navigation.instant and navigation.instant.prefetch.
• If code-heavy content exists: enable content.code.copy, content.code.select, and content.code.annotate.
• If code-heavy content includes API references: standardize on mkdocstrings for generated API docs and keep hand-written task guides alongside generated reference pages.
• If users rely on search often: enable search.highlight and improve heading quality for better snippets.
• If many pages share tab labels (for example Python/JS): enable content.tabs.link.

Completion check: each enabled feature has a documented rationale tied to a user outcome.

Step 5: Standardize Navigation and Search Quality

1. Use explicit nav for critical docs journeys and release-stable ordering.
2. Keep page titles and H1 aligned with search intent language users actually type.
3. Avoid duplicate page titles across sections.
4. Use concise first paragraphs because search snippets often rely on early content.
5. Add cross-links between concept, task, and reference pages.

Completion check: top support queries return a correct page within first search results.

Step 6: Publish and Validate

1. Build the site: uv run zensical build
2. Validate links, code blocks, and navigation paths.
3. Review on desktop and mobile for scanability and heading rhythm.
4. Check that key journeys (new user setup, common task, troubleshooting flow) are uninterrupted.

Completion check: no broken links, no dead-end pages, and all critical journeys are complete.

Completion Checks

• Page taxonomy is intent-based: Learn, Do, Reference, Troubleshoot.
• Every major section has an overview page and next-step links.
• Task docs contain prerequisites, steps, expected results, and recovery guidance.
• Enabled Zensical features are justified and aligned to user needs.
• Code-heavy documentation strategy is explicit, including when and how mkdocstrings is used.
• Navigation and search behavior are verified with real user tasks.
• Site builds cleanly with uv run zensical build.

Output Contract

Return:

1. Proposed docs architecture and navigation map.
2. Zensical feature configuration recommendations with rationale.
3. A prioritized writing backlog (must-have, should-have, nice-to-have).
4. A quality-gate checklist for pre-publish review.

Guardrails

• Do not produce architecture-only docs without concrete task pages.
• Do not bury prerequisites or version constraints below the fold.
• Do not rely only on search; preserve strong navigational paths.
• Do not enable features without documenting the expected user benefit.