/improve-components - Component Quality Improvement

Orchestrates the systematic improvement of CODITECT framework components based on retrospective data, quality standards, and usage patterns.

Usage

/improve-components                        # Full improvement cycle
/improve-components --analyze              # Analysis only (no changes)
/improve-components --report               # Generate report only
/improve-components --priority P0          # Only critical priority
/improve-components --type skill           # Only skills
/improve-components --component <name>     # Single component
/improve-components --dry-run              # Preview changes

Options

Option	Description
--analyze	Run analysis phase only, no modifications
--report	Generate improvement report
--priority P0|P1|P2	Filter by priority level
--type agent|skill|command|hook	Filter by component type
--component <name>	Target specific component
--dry-run	Show what would be changed
--batch <n>	Process n components per run
--auto-commit	Commit changes after improvement

System Prompt

EXECUTION DIRECTIVE: When the user invokes /improve-components, you MUST:

1. Load Retrospective Data

   python3 hooks/session-retrospective.py --analyze-skills

2. Identify Low-Scoring Components

   python3 scripts/skill-pattern-analyzer.py --recommendations

3. Run Analysis Phase

   /agent component-analyzer "Analyze components with <70% success rate"

4. Apply Enhancements (unless --analyze)

   /agent component-enhancer "Apply improvements based on analysis"

5. Validate and Report

   python3 scripts/component-indexer.py
   python3 scripts/update-component-counts.py

Improvement Phases

Phase 1: Discovery

# Get current skill scores
cat context-storage/skill-learnings.json | python3 -c "
import json, sys
d = json.load(sys.stdin)
for skill, data in d.get('skill_history', {}).items():
    rate = data.get('success_rate', 0)
    if rate < 0.7:
        print(f'{skill}: {rate*100:.0f}%')
"

Phase 2: Analysis

For each low-scoring component:

1. Load component file
2. Check against SKILL-QUALITY-STANDARD.md
3. Identify missing sections
4. Calculate quality score
5. Generate recommendations

Phase 3: Enhancement

For each component needing improvement:

1. Add missing Success Output section
2. Add missing Completion Checklist
3. Add When NOT to Use section
4. Add Anti-Patterns table
5. Add Principles footer

Phase 4: Validation

# Re-classify all improved components
for f in <improved-files>; do
    python3 scripts/moe_classifier/classify.py "$f" --update-frontmatter
done

# Re-index
python3 scripts/component-indexer.py

# Update counts
python3 scripts/update-component-counts.py

Phase 5: Reporting

Generate summary with:

• Components analyzed
• Components improved
• Before/after scores
• Remaining gaps
• Next cycle recommendations

Example Output

============================================================
CODITECT Component Improvement Cycle
============================================================

Phase 1: Discovery
  Retrospective data loaded: 100 sessions
  Low-scoring components identified: 25

Phase 2: Analysis
  Analyzing: how.md → 49% (Grade F)
  Analyzing: classify.md → 48% (Grade F)
  Analyzing: work-next.md → 49% (Grade F)
  ...
  Analysis complete: 25 components

Phase 3: Enhancement
  Enhancing: how.md
    + Added Success Output section
    + Added Completion Checklist
    + Added When NOT to Use
    → New score: 72% (Grade C)

  Enhancing: classify.md
    + Added Success Output section
    + Added Anti-Patterns table
    → New score: 71% (Grade C)
  ...
  Enhancement complete: 18 components improved

Phase 4: Validation
  ✓ MoE classification updated: 18/18
  ✓ Markdown lint passed: 18/18
  ✓ Component index updated

Phase 5: Report
============================================================
✅ IMPROVEMENT CYCLE COMPLETE

Summary:
  Components analyzed: 25
  Components improved: 18
  Components skipped: 7 (already compliant)

Score Changes:
  Average before: 52%
  Average after:  74%
  Average gain:   +22%

Top Improvements:
  how.md:        49% → 72% (+23%)
  classify.md:   48% → 71% (+23%)
  work-next.md:  49% → 68% (+19%)

Remaining P0: 0
Remaining P1: 4

Next cycle recommended: 7 days
============================================================

Completion Checklist

Before marking complete:

• Retrospective data loaded
• Low-scoring components identified
• Analysis completed for all targets
• Enhancements applied (or --analyze mode)
• Validation passed
• Report generated

Success Output

✅ IMPROVEMENT CYCLE COMPLETE: /improve-components

Cycle: 2026-01-03
Duration: 15 minutes
Components: 18 improved / 25 analyzed

Average score improvement: +22%
All P0 issues resolved: Yes

Failure Indicators

This command has FAILED if:

• ❌ Retrospective data not available
• ❌ No components identified for improvement
• ❌ Enhancement phase errors
• ❌ Validation failures

When to Use

Use when:

• After retrospective shows low scores
• Before major releases
• Weekly maintenance cycle
• After adding new quality standards

Do NOT use when:

• During active sprint
• When components are being modified
• Without retrospective data

Related Components

Component	Purpose
component-analyzer	Analyzes quality
component-enhancer	Applies improvements
component-improvement skill	Full methodology
SKILL-QUALITY-STANDARD.md	Quality requirements
/optimize-skills	View skill health
/retrospective	Generate retrospective

Principles

This command embodies:

• #1 Recycle, Extend, Re-Use - Improves existing components
• #9 Based on Facts - Uses retrospective metrics
• #10 Research When in Doubt - Follows quality standard

Full Standard: CODITECT-STANDARD-AUTOMATION.md

---

Version: 1.0.0 | Created: 2026-01-03 | Author: CODITECT Team