7.5 KiB
7.5 KiB
name, description, delegates-to
| name | description | delegates-to |
|---|---|---|
| validate:patterns | Validate pattern learning database integrity and generate health reports | autonomous-agent:orchestrator |
Command: /validate:patterns
Validates the pattern learning system across all commands and agents. Ensures patterns are being stored correctly, consistently formatted, and effectively used for improving performance over time.
Purpose
- Validate pattern learning is working across all commands
- Check pattern format consistency and completeness
- Analyze learning effectiveness and trends
- Identify commands that aren't storing patterns
- Generate comprehensive learning analytics
What It Does
1. Command Coverage Validation (10-20 seconds)
- Scan all commands in
commands/directory - Check which commands store patterns vs. utility commands
- Validate pattern storage code presence
- Identify missing pattern integration
2. Agent Learning Validation (10-15 seconds)
- Verify all agents contribute to pattern learning
- Check learning-engine integration points
- Validate agent effectiveness tracking
- Ensure proper handoff protocols
3. Pattern Storage Analysis (15-30 seconds)
- Validate
.claude-patterns/patterns.jsonformat - Check for required fields and data types
- Analyze pattern quality and completeness
- Detect duplicate or corrupted patterns
4. Learning Effectiveness Metrics (10-20 seconds)
- Calculate pattern reuse rates
- Analyze success rates by task type
- Track skill effectiveness over time
- Identify improvement trends
5. Cross-Reference Validation (10-15 seconds)
- Validate skill references in patterns
- Check agent consistency with stored patterns
- Verify tool usage compliance
- Ensure documentation alignment
6. Learning Analytics Report (20-40 seconds)
- Generate comprehensive learning dashboard
- Create visualizations and charts
- Provide improvement recommendations
- Export data for external analysis
Usage
# Basic pattern validation
/validate:patterns
# Include detailed analytics (slower but comprehensive)
/validate:patterns --analytics
# Quick validation skip analytics
/validate:patterns --quick
# Validate specific command or agent
/validate:patterns --filter orchestrator
/validate:patterns --filter release-dev
Output
Terminal Summary (concise)
Pattern Learning Validation Complete ✅
+- Commands Validated: 18/18 (100%)
+- Pattern Storage: Healthy ✅
+- Learning Effectiveness: 94% ✅
+- Issues Found: 0 critical, 2 minor
+- Duration: 1m 45s
📊 Full analytics: .claude/data/reports/validate-patterns-2025-01-15.md
Detailed Report (file)
- Command-by-command validation results
- Pattern storage format validation
- Learning effectiveness metrics with charts
- Agent performance tracking
- Specific issues and fixes needed
- Trend analysis over time
Validation Categories
1. Commands Pattern Storage
Analysis Commands (should store patterns):
/analyze:project✅/analyze:quality✅/validate:fullstack✅/dev:pr-review✅- And 12 more...
Utility Commands (don't store patterns - expected):
/monitor:dashboard- Display only/workspace:reports- File management only
2. Pattern Format Validation
Required fields checked:
{
"task_type": "string",
"context": "object",
"execution": {
"skills_used": "array",
"agents_delegated": "array",
"approach_taken": "string"
},
"outcome": {
"success": "boolean",
"quality_score": "number",
"duration_ms": "number"
},
"reuse_count": "number",
"last_used": "string"
}
3. Learning Effectiveness Metrics
- Pattern Reuse Rate: How often patterns are reused
- Success Rate by Task Type: Performance across different tasks
- Skill Effectiveness: Which skills perform best
- Agent Performance: Agent reliability and speed
- Improvement Trend: Learning progress over time
Integration
The /validate-patterns command integrates with:
- learning-engine agent: Validates pattern capture and storage
- pattern-learning skill: Validates pattern format and structure
- performance-analytics skill: Generates learning metrics
- orchestrator: Uses validation to improve pattern selection
Expected Validation Results
Successful Validation (what you should see)
- 18/18 commands validated
- All analysis commands storing patterns
- Pattern format consistent
- Learning effectiveness > 80%
- No critical issues
Common Issues and Fixes
-
Missing Pattern Storage
- Issue: Command not storing patterns when it should
- Fix: Add pattern learning integration
-
Format Inconsistencies
- Issue: Missing required fields in patterns
- Fix: Update pattern generation code
-
Low Reuse Rate
- Issue: Patterns not being reused effectively
- Fix: Improve pattern matching algorithm
-
Storage Location Issues
- Issue: Patterns not in
.claude-patterns/ - Fix: Update storage path configuration
- Issue: Patterns not in
Analytics Dashboard
When using --analytics flag, generates:
Learning Metrics
- Total patterns stored: 247
- Average reuse count: 3.2
- Success rate: 89%
- Most reused pattern: "refactor-auth-module" (12 times)
Skill Performance
Top Performing Skills:
1. code-analysis (94% success, 45 uses)
2. quality-standards (91% success, 38 uses)
3. pattern-learning (89% success, 52 uses)
Agent Performance
Agent Reliability:
1. orchestrator: 96% success
2. code-analyzer: 94% success
3. quality-controller: 92% success
Usage Examples
Example 1: Basic Validation
User: /validate:patterns
System: ✅ Pattern learning system healthy
Commands storing patterns: 16/16
Pattern format: Valid
Learning effectiveness: 91%
Example 2: With Analytics
User: /validate:patterns --analytics
System: 📊 Generated comprehensive analytics
Learning trends: Improving (+12% over 30 days)
Top skill: code-analysis (95% success)
Recommendation: Increase pattern reuse threshold
Example 3: Filter Validation
User: /validate:patterns --filter orchestrator
System: ✅ Orchestrator pattern integration validated
Patterns contributed: 89
Effectiveness score: 96%
Integration quality: Excellent
When to Use
Run /validate:patterns when:
- After implementing new commands or agents
- Suspecting pattern learning issues
- Regular system health checks
- Before major releases
- Analyzing learning effectiveness
Automation
The orchestrator can automatically run /validate:patterns:
- Every 50 tasks to ensure system health
- When learning effectiveness drops below 75%
- After adding new commands or agents
- During system diagnostics
Troubleshooting
Common Validation Failures
-
Pattern Database Missing
Error: .claude-patterns/patterns.json not found Fix: Run /learn:init to initialize -
Permission Issues
Error: Cannot read pattern database Fix: Check file permissions in .claude-patterns/ -
Corrupted Patterns
Error: Invalid JSON in patterns Fix: Manual repair or reset patterns
Related Commands
/learn:init- Initialize pattern learning system/analyze:project- Analyze project and learn patterns/analyze:quality- Check overall system quality