15 KiB
description, allowed-tools
| description | allowed-tools |
|---|---|
| Documentation review system help - usage guide and command reference | Bash(cat:*) |
📚 Documentation Review System - Help Guide
Version: Phase 3 (Modular Architecture) Last Updated: 2025-10-24
🎯 What Is This?
The /ck:doc-review system is a modular documentation update toolkit that
helps you:
✅ Keep documentation in sync with code changes ✅ Update README, CLAUDE.md, CHANGELOG consistently ✅ Maintain SDD (Specification-Driven Development) artifacts ✅ Validate documentation quality before commits ✅ Save 88% tokens compared to manual updates
🏗️ Architecture Overview
┌─────────────────────────────────────────┐
│ /ck:doc-review (Orchestrator) │
│ • Analyzes changes │
│ • Asks what to update │
│ • Delegates to sub-commands │
│ • Shows summary │
└─────────────────────────────────────────┘
│
┌──────┼──────┬──────┬──────┐
▼ ▼ ▼ ▼ ▼
┌────────┐ │ │ │ │ │ │ └─────────┐
│:analyze│ │:core│ │:sdd│ │:qa│ │:help │
│ │ │ │ │ │ │ │ │ │
│ 0 edits│ │3 files││SDD │ │Check││This!│
└────────┘ └─────┘ └────┘ └────┘ └──────┘
📋 Command Reference
/ck:doc-review/help (This Command)
Purpose: Show help and usage guide Token Cost: ~200 tokens (one-time) Files Modified: None Use When: First time using system, or need reminder
/ck:doc-review/analyze
Purpose: Analyze documentation needs without making changes Token Cost: ~600-800 tokens Files Modified: None Execution Time: < 1 second
What It Does:
- ✅ Detects project documentation principles
- ✅ Analyzes current documentation structure
- ✅ Identifies what needs updating (from git history)
- ✅ Provides actionable recommendations
- ❌ Makes zero changes
When to Use:
- "What docs need updating?"
- Before starting documentation work
- Quick health check
- Understanding project doc structure
Example:
/ck:doc-review/analyze
→ Output:
- 61 .md files found
- Recent changes: 5 files in src/
- Recommendations: Update README, CLAUDE, specs/005/
- Execution: 0.8s
/ck:doc-review/core
Purpose: Update core documentation files Token Cost: ~1,200-1,500 tokens Files Modified: README.md, CLAUDE.md, CHANGELOG.md Execution Time: 15-30 seconds
What It Does:
- ✅ Updates README.md (features, usage, installation)
- ✅ Updates CLAUDE.md (AI context, module docs)
- ✅ Updates CHANGELOG.md (version entry)
- ✅ Uses professional templates
- ✅ Validates changes
When to Use:
- Just added a new feature
- Need to document new API/module
- Quick documentation update
- Before sharing project
Example:
/ck:doc-review/core "ApstraWorkModule implementation"
→ Updates:
- README.md: Added usage examples
- CLAUDE.md: Added module context
- CHANGELOG.md: Created v1.2.0 entry
- Time: 18s
Templates Provided:
- README feature section
- CLAUDE module documentation
- CHANGELOG version entry
/ck:doc-review/sdd
Purpose: Update SDD (Specification-Driven Development) artifacts Token Cost: ~1,500-1,800 tokens Files Modified: specs//spec.md, plan.md, tasks.md, contracts/.md Execution Time: 30-60 seconds
What It Does:
- ✅ Marks requirements (FR-*) as completed in spec.md
- ✅ Updates implementation plan and phase status in plan.md
- ✅ Marks tasks complete and adds notes in tasks.md
- ✅ Updates design contracts and API specs
- ✅ Validates SDD ↔ code consistency
When to Use:
- Completed an implementation phase
- Need to sync specs with code
- Mark requirements as done
- Document architecture decisions
Example:
/ck:doc-review/sdd "Phase 2 completion"
→ Updates:
- spec.md: Marked FR-005-1, FR-005-2 complete
- plan.md: Phase 2 status = ✅ Completed
- tasks.md: 15 tasks marked done
- Consistency: 100%
- Time: 42s
SDD Consistency Checks:
- Requirements → Implementation alignment
- Plan → Tasks alignment
- Contracts → Code alignment
- Traceability validation
/ck:doc-review/qa
Purpose: Quality assurance - validate documentation Token Cost: ~1,800-2,000 tokens Files Modified: None Execution Time: 10-20 seconds
What It Does:
- ✅ Validates markdown links (finds broken links)
- ✅ Checks file:line references (path/to/file.py:123)
- ✅ Checks terminology consistency
- ✅ Validates version number consistency
- ✅ Checks SDD requirement traceability
- ✅ Basic code example syntax validation
- ✅ Completeness check (required sections)
- ✅ Generates scored report (0-100)
When to Use:
- Before committing documentation changes
- After major documentation updates
- Find and fix broken links
- Ensure documentation quality
Example:
/ck:doc-review/qa
→ Results:
- Valid links: 156
- Broken links: 3
- Version mismatches: 1
- Overall score: 87/100
- Time: 12s
→ Recommendations:
1. Fix broken link in README.md:45
2. Update version in pyproject.toml
QA Categories:
- Link Validation (20 points)
- Cross-Reference Validation (15 points)
- Terminology Consistency (15 points)
- Version Consistency (15 points)
- SDD Consistency (15 points)
- Code Example Quality (10 points)
- Completeness (10 points)
/ck:doc-review (Main Orchestrator)
Purpose: Smart orchestration - analyze and delegate Token Cost: ~900-1,200 tokens (orchestrator only) + selected sub-commands Files Modified: Depends on user selection Execution Time: Variable (depends on selection)
What It Does:
- ✅ Runs comprehensive analysis (via doc-analyzer.sh)
- ✅ Shows analysis summary
- ✅ Asks user which areas to update
- ✅ Invokes selected sub-commands automatically
- ✅ Aggregates and shows summary
When to Use:
- Not sure what needs updating
- Want guided workflow
- Comprehensive documentation update
- First time using the system
Example:
/ck:doc-review "feature 005 implementation"
→ Analysis shown (principles, structure, changes, metrics)
→ User asked:
Which areas to update?
□ Core Files (README, CLAUDE, CHANGELOG)
□ SDD Artifacts (spec, plan, tasks)
□ Run QA Only
□ Full Update
→ User selects "Full Update"
→ Executes:
1. /ck:doc-review/core "feature 005"
2. /ck:doc-review/sdd "feature 005"
3. /ck:doc-review/qa
→ Aggregated summary:
- 8 files updated
- QA score: 92/100
- Total time: 78s
🎓 Usage Patterns
Pattern 1: Quick README Update (Most Common)
# Scenario: Added new feature, need docs
/ck:doc-review/core "new feature name"
# Time: 20s
# Token: 1.2K
# Updates: README, CLAUDE, CHANGELOG
Pattern 2: Analysis First (Recommended for New Users)
# Step 1: Understand what needs updating
/ck:doc-review/analyze
# Step 2: Update based on recommendations
/ck:doc-review/core "feature X"
# Time: 1s + 20s = 21s
# Tokens: 600 + 1.2K = 1.8K
Pattern 3: Full Update After Major Work
# Completed major implementation
/ck:doc-review "Phase 2 complete"
→ Select "Full Update"
→ All docs updated + QA validated
# Time: 60-90s
# Tokens: 3-4K
# Updates: Everything
Pattern 4: QA Before Commit (Best Practice)
# Before git commit
/ck:doc-review/qa
→ Fix any issues found
→ Re-run QA
→ Commit with confidence
# Time: 10-15s per run
# Tokens: 1.8K
Pattern 5: SDD Update After Phase
# Completed implementation phase
/ck:doc-review/sdd "Phase 2"
# Time: 40-50s
# Tokens: 1.5K
# Updates: All SDD artifacts
🗂️ File Structure Reference
Command Files Location
~/.claude/commands/ck/
├── doc-review.md # Main orchestrator
├── doc-review:analyze.md # Analysis only
├── doc-review:core.md # Core files update
├── doc-review:sdd.md # SDD artifacts update
├── doc-review:qa.md # Quality validation
├── doc-review:help.md # This help file
└── doc-categories.json # Pattern configuration
Supporting Tools
~/.claude/tools/
└── doc-analyzer.sh # External analysis script
├── principles # Extract doc principles
├── structure # Analyze doc structure
├── categorize # Categorize files
├── impact # Assess change impact
├── metrics # Generate metrics
└── all # Run all analyses
Documentation
docs/
├── ck-doc-review-analysis.md # Architecture analysis
├── ck-doc-review-phase1-improvements.md # Phase 1 docs
├── ck-doc-review-phase2-improvements.md # Phase 2 docs
└── ck-doc-review-phase3-improvements.md # Phase 3 docs (current)
📊 Token Budgeting
Per-Command Token Costs
| Command | Tokens | Use Case |
|---|---|---|
:help |
~200 | One-time reference |
:analyze |
~600-800 | Analysis only |
:core |
~1.2-1.5K | Quick update |
:sdd |
~1.5-1.8K | SDD update |
:qa |
~1.8-2K | Quality check |
doc-review |
~900-1.2K + sub-commands | Full orchestration |
Monthly Token Budget Examples
Light Usage (10 updates/month):
10x :core updates: 12-15K tokens
2x :qa checks: 3.6-4K tokens
1x :help reference: 200 tokens
Total: ~16-19K tokens/month
Medium Usage (20 updates/month):
15x :core updates: 18-22.5K tokens
4x :sdd updates: 6-7.2K tokens
4x :qa checks: 7.2-8K tokens
1x :help reference: 200 tokens
Total: ~31-38K tokens/month
Heavy Usage (40 updates/month):
25x :core updates: 30-37.5K tokens
8x :sdd updates: 12-14.4K tokens
8x :qa checks: 14.4-16K tokens
4x orchestrated: 12-16K tokens
1x :help reference: 200 tokens
Total: ~69-84K tokens/month
Original System (Same Usage):
40x monolithic commands: 400-480K tokens/month
Savings with Phase 3: 83-93%
🔧 Troubleshooting
"Command not found: /ck:doc-review/core"
Cause: Sub-command file not installed Solution:
# Check if files exist
ls ~/.claude/commands/ck/doc-review*.md
# Should see all sub-command files
# If missing, reinstall from docs
"Analysis fails with error"
Cause: doc-analyzer.sh not executable or not found Solution:
# Check if tool exists
ls -l ~/.claude/tools/doc-analyzer.sh
# Make executable
chmod +x ~/.claude/tools/doc-analyzer.sh
# Test manually
~/.claude/tools/doc-analyzer.sh help
"QA reports many broken links"
Cause: Links broken after file reorganization Solution:
# Run QA to identify broken links
/ck:doc-review/qa
# Fix links in reported files
# Re-run QA to verify
/ck:doc-review/qa
"SDD consistency check fails"
Cause: Requirements/tasks out of sync with code Solution:
# Update SDD artifacts
/ck:doc-review/sdd "current feature"
# Manually review any remaining inconsistencies
# Update as needed
💡 Tips & Best Practices
1. Run :analyze First (New Projects)
# When starting with a new project
/ck:doc-review/analyze
→ Understand doc structure
→ Identify gaps
→ Plan updates
2. Use :core for Quick Updates
# Daily workflow: added feature, need docs
/ck:doc-review/core "feature name"
→ Fast (20s)
→ Low token cost (1.2K)
→ Updates essentials
3. Run :qa Before Commits
# Git workflow
git add .
/ck:doc-review/qa # Check doc quality
→ Fix any issues
git commit -m "docs: ..."
4. Use Orchestrator When Unsure
# Not sure what needs updating?
/ck:doc-review "what changed"
→ Guided workflow
→ Smart recommendations
→ You choose scope
5. Batch SDD Updates
# Don't update SDD after every task
# Batch at phase boundaries
/ck:doc-review/sdd "Phase 2 complete"
→ More efficient
→ Better overview
→ Cleaner commits
🎯 Decision Tree: Which Command to Use?
Start
│
├─ Need help? → /ck:doc-review/help
│
├─ Just want to see what needs updating?
│ └─ /ck:doc-review/analyze
│
├─ Added new feature/module?
│ └─ /ck:doc-review/core "feature name"
│
├─ Completed implementation phase?
│ └─ /ck:doc-review/sdd "phase name"
│
├─ About to commit documentation?
│ └─ /ck:doc-review/qa
│
├─ Not sure what to do?
│ └─ /ck:doc-review "what changed"
│
└─ Want comprehensive update?
└─ /ck:doc-review "scope" → Select "Full Update"
📚 Further Reading
Architecture Deep Dive:
docs/ck-doc-review-analysis.md- Initial architecture analysisdocs/ck-doc-review-phase3-improvements.md- Current system design
Phase Evolution:
- Phase 1: Progressive disclosure + patterns config
- Phase 2: External tool architecture
- Phase 3: Full modularization (current)
External Tools:
~/.claude/tools/doc-analyzer.sh help- Tool documentation- Run
doc-analyzer.sh allto see full analysis output
🆘 Getting Help
Within the System
/ck:doc-review/help # This guide
~/.claude/tools/doc-analyzer.sh help # Tool help
Documentation Files
cat docs/ck-doc-review-phase3-improvements.md # Full system docs
Quick Reference Card
# Analysis
/ck:doc-review/analyze
# Updates
/ck:doc-review/core "feature" # Quick
/ck:doc-review/sdd "phase" # SDD
/ck:doc-review "scope" # Full
# Validation
/ck:doc-review/qa
# Help
/ck:doc-review/help
✨ Key Takeaways
- Modular = Efficient - Use focused commands, not monolithic tool
- Analyze First - Understand before updating
- Core for Speed - Most common use case
- QA Before Commit - Catch issues early
- Orchestrate When Unsure - Guided workflow helps
Last Updated: 2025-10-24 Version: Phase 3 (Modular Architecture) System Status: Production Ready ✅