Documentation

Technical documentation patterns, structure, maintenance, and avoiding common documentation failures.

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "Documentation" with this command: npx skills add ivangdavila/documentation

Structure Hierarchy

  • README: what it is, how to install, quick example — 5 minutes to first success
  • Getting Started: guided tutorial for beginners — one complete workflow
  • Guides: task-oriented ("How to X") — goal-focused, not feature-focused
  • Reference: exhaustive API/CLI docs — complete but not for learning
  • Troubleshooting: common errors with solutions — search-optimized

README Essentials

  1. One-sentence description — what problem it solves
  2. Installation — copy-paste command that works
  3. Quick start — minimal example that actually runs
  4. Link to full docs — don't cram everything in README

Missing any of these = users bounce before trying.

Code Examples

  • Every example must be tested — untested examples rot within months
  • Show complete runnable code, not fragments — users copy-paste
  • Include expected output — confirms they did it right
  • Bad: client.query(...) / Good: full script with imports, setup, and output
  • Version-pin examples: npm install package@2.1.0 not npm install package

API Documentation

  • Every endpoint needs: method, path, parameters, request body, response, error codes
  • Show real request/response bodies — not just schemas
  • Include authentication in every example — most common missing piece
  • Document rate limits and pagination upfront — not buried in footnotes
  • Error responses need as much detail as success responses

What Gets Outdated

  • Screenshots — UI changes, screenshots don't
  • Version numbers — hardcoded versions become wrong
  • Links — external sites move, break constantly
  • "Current" anything — write timelessly or add review dates
  • Feature flags and experimental warnings — often forgotten after GA

Maintenance Patterns

  • Docs live next to code — same repo, same PR. Separate repos drift
  • CI checks for broken links — markdown-link-check or equivalent
  • Runnable examples as tests — if example breaks, build fails
  • Review date in docs: "Last verified: 2024-01" — signals freshness
  • Delete aggressively — outdated docs worse than no docs

Common Failures

  • Documenting implementation, not usage — users don't care how it works internally
  • Assuming context — define acronyms, link prerequisites
  • Wall of text — use headings, bullets, code blocks liberally
  • "See X for more info" without link — friction kills follow-through
  • Changelog as documentation — changes ≠ how to use current version

Writing Style

  • Imperative mood: "Run the command" not "You can run the command"
  • Second person: "you" not "the user"
  • Present tense: "This returns X" not "This will return X"
  • Short sentences — one idea per sentence
  • Active voice: "The function returns X" not "X is returned by the function"

Searchability

  • Use words users search for — not internal jargon
  • Error messages verbatim in troubleshooting — users paste exact errors
  • Multiple ways to describe same thing — alias common variations
  • H2/H3 headings are SEO — match user queries
  • Avoid clever titles — "Getting Started" beats "Your Journey Begins"

Versioned Documentation

  • Major versions need separate docs — v1 users shouldn't see v2 docs
  • Migration guides between versions — step-by-step, not just changelog
  • Default to latest stable, link to older versions
  • Mark deprecated features clearly — don't just remove
  • URL structure: /docs/v2/ not query params

README Anti-patterns

  • Badge spam — 15 badges before content
  • Massive feature lists — save for marketing page
  • No installation instructions — assuming everyone knows
  • Screenshots without context — what am I looking at?
  • License-only README — legal compliance ≠ documentation

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.