Document Skill for Claude
Generate @moduledoc/@doc for tested Elixir features; may update their README section or ADR. Not for docs lookup, documentation audits/reviews, or capturing standalone decisions.
Install
one command, takes just this skill from the repositorynpx skills add https://github.com/oliver-kriska/claude-elixir-phoenix --skill documentWhat comes with it
9 577 bytes besides the instructionThe instruction itself
9 sections, as written by the authorDocument
Generate documentation for newly implemented features.
Usage
/phx:document .claude/plans/magic-link-auth/plan.md
/phx:document magic link authentication
/phx:document # Auto-detect from recent plan
Iron Laws
- Never remove existing documentation — Existing docs may reflect design intent that isn't obvious from code alone; update rather than replace
- @moduledoc on every public module — Undocumented modules accumulate quickly and create onboarding friction for new team members
- ADRs capture the "why", not the "what" — Code shows what was built; ADRs explain why this approach was chosen over alternatives
- Match @doc to function's public API — Document parameters, return values, and edge cases; callers shouldn't need to read the implementation
- DO NOT add @doc to untested code — documentation implies a stable contract; document only after tests confirm the function behaves as described
What Gets Documented
| Output | Description |
|--------|-------------|
| @moduledoc | For new modules missing documentation |
| @doc | For public functions without docs |
| README section | For user-facing features |
| ADR | For significant architectural decisions |
Workflow
Step 0: Pre-check (avoid no-op runs)
Run git diff --name-only HEAD~5 | grep '\.ex$' | head -20 to check for new .ex files.
If NO new .ex files were added (only modifications), skip the full
audit and report: "No new modules — documentation coverage unchanged."
This prevents 35-message analysis sessions that conclude "PASS" with
zero output (confirmed: session bb0a0454 wasted ~2K tokens on no-op).
- Identify new modules from recent commits or plan file
- Check documentation coverage (
@moduledoc,@doc) - Generate missing docs using templates
- Add README section if user-facing feature
- Create ADR if architectural decision was made
- Write report to
.claude/plans/{slug}/reviews/{feature}-docs.md
When to Generate ADRs
| Trigger | Create ADR |
|---------|-----------|
| New external dependency | Yes |
| New database table | Maybe (if schema non-obvious) |
| New OTP process | Yes (explain why process needed) |
| New context | Maybe (if boundaries non-obvious) |
| New auth mechanism | Yes |
| Performance optimization | Yes |
Integration with Workflow
/phx:plan → /phx:work → /phx:review
↓
/phx:document ← YOU ARE HERE (optional, suggested after review passes)
References
${CLAUDE_SKILL_DIR}/references/doc-templates.md— @moduledoc, @doc, README, ADR templates${CLAUDE_SKILL_DIR}/references/output-format.md— Documentation report format${CLAUDE_SKILL_DIR}/references/doc-best-practices.md— Elixir documentation best practices${CLAUDE_SKILL_DIR}/references/documentation-patterns.md— Detailed documentation patterns
Other skills for the same job
different authors, same section of the catalogue
Create time-boxed technical spike documents for researching and resolving critical development decisions before implementation.
Automatically convert Confluence specification documents into structured Jira backlogs with Epics and implementation tickets. When an agent needs to: (1) Create Jira tickets from a Confluence page, (2) Generate a backlog from a specification, (3) Break down a spec into implementation tasks, or (4) Convert requirements into Jira issues. Handles reading Confluence pages, analyzing specifications, creating Epics with proper structure, and generating detailed implementation tickets linked to the Epic.
Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.
Guidelines for clinical decision support (CDS) documents: biomarker-stratified cohort analyses and GRADE-graded treatment reports. Covers structure, executive summaries, evidence grading (1A–2C), stats (HR, CI, survival), and biomarker integration. Use for pharma research docs, clinical guidelines, regulatory submissions.
Generate professional clinical decision support (CDS) documents for pharmaceutical and clinical research settings, including patient cohort analyses (biomarker-stratified with outcomes) and treatment recommendation reports (evidence-based guidelines with decision algorithms). Supports GRADE evidence grading, statistical analysis (hazard ratios, survival curves, waterfall plots), biomarker integration, and regulatory compliance. Outputs publication-ready LaTeX/PDF format optimized for drug development, clinical research, and evidence synthesis.
Guides through Trail of Bits' 5-step secure development workflow. Runs Slither scans, checks special features (upgradeability/ERC conformance/token integration), generates visual security diagrams, helps document security properties for fuzzing/verification, and reviews manual security areas.
Document architecture decisions with ADR (Architecture Decision Records). Use when making significant technical decisions, choosing between alternatives, or when onboarding needs context on past decisions.
Plan-approval workflow patterns for user control over AI actions in Claude Code Waypoint Plugin. Use when planning complex changes, need user approval before execution, want to prevent mistakes, or need to document proposed changes. Covers plan creation, approval checkpoints, plan deviation tracking, revision management, and learning from approved/rejected plans.
How to use it
Take oliver-kriska/document from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.