Construct-AI-primary Documentation Quality Check
Overview
This skill defines the validation checklist for all documentation. Good documentation is accurate, complete, clear, structured, and maintained. Poor documentation wastes time and causes errors.
When to Use
- Before delivering any documentation to another agent or stakeholder
- Before publishing documentation to shared repositories
- After making changes to existing documentation
- When documentation is reviewed as part of quality gates
- Don't use when: Documentation is exploratory draft not yet for distribution
Core Procedures
Step 1: Accuracy Check
- All technical information is correct and current
- Examples produce stated results when followed
- References to other documents are valid and point to correct locations
- No outdated or deprecated information without clear marking
- Diagrams, screenshots, and visual aids accurately represent the system
Step 2: Completeness Check
- All intended topics are covered
- Prerequisites and dependencies documented
- Scope and limitations clearly stated
- Audience-appropriate detail level (not too shallow, not too deep)
- Known gaps or TODOs are explicitly identified
Step 3: Clarity Check
- Writing is clear and free of jargon without explanation
- Structure is logical and easy to navigate
- Headings and sections are descriptive
- Links work and point to correct destinations
- Table of contents matches actual structure (if present)
- No broken formatting or display issues
Step 4: Consistency Check
- Terminology used consistently throughout
- Formatting follows project/style standards
- Cross-references are accurate and bidirectional
- Code examples follow coding-standards
- Version information matches actual document version
Step 5: Actionability Check
- Procedures have clear step-by-step instructions
- Expected outcomes described for each procedure
- Troubleshooting guidance provided for common failures
- Contact or escalation path included where relevant
- Document has author, date, and version metadata
Quality Checklist
- ALL accuracy checks pass
- ALL completeness checks pass
- ALL clarity checks pass
- ALL consistency checks pass
- ALL actionability checks pass
- Documentation reviewed by second party if high-impact
Error Handling
- Error: Technical inaccuracy discovered Response: Fix immediately, re-run all accuracy checks, note change in version history
- Error: Documentation is incomplete Response: Complete missing sections before delivery, or clearly mark as partial with expected completion
- Error: Examples don't work as documented Response: Fix example or fix documentation, whichever is actually correct
Cross-Team Integration
Related Skills: output-validation-checklist, source-verification, taxonomy-standards, feedback-integration Used By: Documenters, ALL agents producing documentation, especially technical writers and API documenters