"Apply Diataxis framework to organize workflow output into documentation directories. Use to guide skill output placement: inline by default unless user requests file; rationale-heavy content to docs/architecture/; lookup tables and reference guides to docs/reference/. Avoid standalone analysis summaries in docs root. Composable into any analysis or decision-making skill."
Install
npx skillscat add squirrel289/pax/organizing-documents-diataxis Install via the SkillsCat registry.
Aspect: Organizing Documents (Diataxis)
Reusable aspect providing consistent guidance for placing skill-generated documentation into Diataxis-aligned directories.
Purpose
Standardize where workflow outputs and analysis results should be placed in the documentation hierarchy, based on the Diátaxis framework categorization.
When to Use This Aspect
Compose this aspect into any skill that:
- Produces analysis records or comparative decision results
- May generate documentation files
- Needs guidance on organizing outputs by type (explanation, reference, etc.)
- Should follow PAX documentation conventions
The Diátaxis Framework
PAX documentation uses the Diátaxis framework to organize content:
| Type | Purpose | Location | Style | Examples |
|---|---|---|---|---|
| Tutorials | Learning-oriented, step-by-step | docs/guides/ |
Hands-on, narrative | Getting started guides |
| How-to Guides | Task-oriented, problem-solving | docs/guides/ |
Recipe format, context-specific | Common workflows, checklists |
| Reference | Information-oriented, complete | docs/reference/ |
Dry, precise, lookup-oriented | Decision trees, quick refs, schemas, matrices |
| Explanation | Understanding-oriented, rationale | docs/architecture/ |
Conceptual, decision drivers, theory | Analysis records, design decisions, reports |
Output Placement Rules
Rule 1: Default to Inline Output
Unless the user explicitly requests a file:
- Provide analysis results, recommendations, and decision records inline in the chat
- Do not automatically create files in
docs/ - This respects user intent and AI cost constraints
Example:
User: "Compare these tools using comparative-analysis"
→ Provide ranking + recommendation inline
→ Do not create docs/architecture/comparison-record.md automaticallyRule 2: Rationale-Heavy Records → docs/architecture/
When creating files, place explanation-oriented outputs here:
- Comparative decision records with full rationale
- Analysis reports with decision drivers
- Architecture recommendations with justification
- Evaluation summaries with evidence traces
Naming: Use descriptive names with context
- ✅ Good:
pr-workflow-analysis-2026-02.md,comparing-parsing-approaches.md - ❌ Bad:
analysis.md,comparison.md,report.md
Format: Include:
- Decision context (what was being decided)
- Evaluation criteria and evidence
- Ranked options with score breakdowns
- Selected recommendation with rationale
- Evidence gaps and follow-up actions
Rule 3: Reference Tables and Decision Trees → docs/reference/
Place information-oriented lookup content here:
- Quick reference tables (tools, features, platform matrix)
- Decision trees (routing, escalation, workflow selection)
- Schema summaries and field catalogs
- Index and lookup tables for complex information
Naming: Use short, lookup-friendly names
- ✅ Good:
tool-feature-matrix.md,workflow-decision-tree.md,api-quick-ref.md - ❌ Bad:
tools-comparison-analysis.md(too detailed for reference)
Format:
- Scannable tables or short descriptions
- No lengthy explanations (move rationale to
docs/architecture/) - Cross-references to fuller explanations in
docs/architecture/if needed
Rule 4: Avoid Docs Root Analysis Summaries
Do not create ephemeral analysis files in the docs/ root:
- ❌
docs/ANALYSIS_SUMMARY.md(ephemeral, clutters docs root) - ❌
docs/SKILL_REVIEW_ASSESSMENT.md(should be in architecture/ or omitted) - ❌
docs/DECISION_SUMMARY.md(belongs in architecture/ or inline)
Why: Analysis summaries are output artifacts, not documentation. The docs/ root should contain durable, user-facing documentation.
Decision Flow
User asks for analysis/comparison
↓
1. Generate output inline by default
↓
2. User asks "save this to a file"?
├─ NO → done (inline output sufficient)
└─ YES → proceed to step 3
↓
3. Is the content primarily rationale/explanation?
├─ YES → docs/architecture/{descriptive-name}.md
└─ NO → proceed to step 4
↓
4. Is the content a lookup table or reference guide?
├─ YES → docs/reference/{lookup-name}.md
└─ NO → reconsider if file output needed; prefer inlineExamples
Example 1: Skill Comparison Analysis
User request: "Compare skills X, Y, Z using comparative-analysis and save results"
Output placement:
- Inline: Ranked list with top recommendation
- If file requested:
- Summary table →
docs/reference/skill-comparison-matrix.md(lookup table) - Full analysis with evidence →
docs/architecture/skill-selection-analysis.md(rationale)
- Summary table →
Example 2: Tool Evaluation Record
User request: "Evaluate these build tools with scoring and recommendation"
Output placement:
- Inline: Recommendation and top option summary
- If file requested:
- Full comparative decision record →
docs/architecture/build-tool-evaluation-2026-02.md - Quick reference table →
docs/reference/build-tool-feature-matrix.md(optional, if useful for future lookups)
- Full comparative decision record →
Example 3: Workflow Decision
User request: "Decide on PR merge strategy"
Output placement:
- Inline: Selected strategy and decision rationale
- If file requested:
- Decision record →
docs/architecture/pr-merge-strategy-decision.md - Decision tree for future use →
docs/reference/pr-strategy-decision-tree.md
- Decision record →
Composing This Aspect
When a workflow skill uses this aspect, document it in the skill's section:
## Output Placement and File Organization (Diataxis)
This skill uses the `aspect-organizing-documents-diataxis` aspect to guide output placement:
- Default to inline output unless the user requests a file.
- Place rationale-heavy records in `docs/architecture/`.
- Place lookup tables or decision trees in `docs/reference/`.
- Avoid creating standalone analysis summaries in the docs root.
See [[skills/aspects/organizing-documents-diataxis/SKILL.md]] for full guidance.Then, when generating outputs:
- Follow the output placement rules above
- Name files descriptively (context + date if multiple versions)
- Structure content according to category (explanation vs. reference)
- Always provide inline summary; ask before creating files
Related Documentation
- Diátaxis Framework: https://diataxis.fr/
- File Organization: [[docs/reference/FILE_ORGANIZATION.md]]
- Naming Conventions: [[docs/conventions/NAMING_CONVENTIONS.md]]
- Frontmatter Specification: [[docs/conventions/FRONTMATTER_SPECIFICATION.md]]