midjourney-cc-skill: Command for Claude Code

Show Knowledge

Display the current state of learned knowledge, optionally filtered by category.

Instructions

1. Verify database access. Run SELECT COUNT(*) FROM sessions via sqlite-simple MCP. If the query fails, tell the user: "Database not available. Run claude mcp add sqlite-simple -- npx @anthropic-ai/sqlite-simple-mcp mydatabase.db then restart Claude Code." Do not proceed without database access.

2. Parse the optional category filter from user input. Valid categories:

   • lighting, materials, forms, parameters, prompt-structure, color, composition, style, mood, v7, failure-modes, keywords
   • If no category specified, show an overview of all.

3. If showing overview (no category filter):

   SELECT category, COUNT(*) as pattern_count,
     SUM(CASE WHEN confidence = 'high' THEN 1 ELSE 0 END) as high_confidence,
     SUM(CASE WHEN confidence = 'medium' THEN 1 ELSE 0 END) as medium_confidence,
     SUM(CASE WHEN confidence = 'low' THEN 1 ELSE 0 END) as low_confidence
   FROM patterns WHERE is_active = 1
   GROUP BY category
   ORDER BY pattern_count DESC
   

   Also show:

   SELECT COUNT(*) as total_sessions, SUM(total_iterations) as total_iterations,
     SUM(CASE WHEN status = 'success' THEN 1 ELSE 0 END) as successful_sessions
   FROM sessions
   

   And:

   SELECT COUNT(*) as total_keywords FROM keyword_effectiveness
   

4. If showing a specific category:

   SELECT * FROM v_pattern_summary WHERE category = ? AND is_active = 1 ORDER BY confidence DESC, success_rate DESC
   

   For each pattern, display:

   • Problem / Solution
   • Example bad / Example good
   • Confidence level and success rate
   • Evidence count (supporting vs contradicting)
   • When it was discovered and last validated

5. If category is 'keywords':

   SELECT * FROM keyword_effectiveness ORDER BY effectiveness DESC, times_used DESC
   

6. If category is 'failure-modes': Read and display knowledge/failure-modes.md.