Show Learnings

Display all learning points that have been automatically tracked by the self-improvement system.

Your Task

Execute the view-learnings script to show tracked learnings.

bash ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/view-learnings.sh

What This Shows

The system automatically tracks learning points when specific topics or issues come up in conversations:

Learning Categories:

• Security: SQL injection, XSS, authentication, etc.
• Best Practices: Testing, validation, error handling
• Code Quality: Structure, maintainability, performance
• Communication: Clarity, completeness, user focus

Learning Information:

• Key: Learning identifier
• Text: The actual learning point
• Learned at: When first tracked
• Reinforced count: How many times it came up again
• Last reinforced: Most recent mention

Example Output

🔥 sql_injection_discussed
   SQL injection topic came up - ensure parameterized queries used
   Learned: 2025-01-10T09:00:00
   Reinforced: 5 times (last: 2025-01-20T15:30:00)

⭐ missing_validation_pattern
   Input validation missing - add validation before business logic
   Learned: 2025-01-12T11:20:00
   Reinforced: 2 times (last: 2025-01-19T14:10:00)

💡 test_coverage_low
   Tests not provided with code - always include test examples
   Learned: 2025-01-18T16:45:00

Learning Indicators

• 🔥 Highly Reinforced (3+ times): Critical learning that keeps coming up
• ⭐ Reinforced (1+ times): Important learning mentioned multiple times
• 💡 New (0 times): Recently learned, not yet reinforced

Usage

/show-learnings

Applying Learnings

When you see highly reinforced learnings:

1. Internalize the pattern: Make it automatic
2. Proactive application: Use it before issues arise
3. Create checks: Add to mental checklist
4. Reduce count: Goal is to never trigger the learning again

Example Application:

Learning: "SQL injection - use parameterized queries"
→ Before writing ANY database query:
  ✓ Check if parameterized
  ✓ Never concatenate user input
  ✓ Use ORM or prepared statements

Reinforcement Tracking

Why track reinforcement?

• Shows if learning is being applied
• High reinforcement = pattern not internalized yet
• Low reinforcement = successfully learning
• Zero reinforcement after learning = mastery!

Goal: Every highly-reinforced learning should eventually stop being reinforced because you've internalized it and stopped making that mistake.

Related Commands

• /show-patterns - View tracked patterns
• /show-metrics - View conversation statistics
• /review-my-work - Trigger manual review
• /quality-check - Quick quality assessment

---

Learnings are automatically tracked when specific topics arise to build a knowledge base for continuous improvement.