ToruAI Capture Learning
User wants to save something to megg knowledge.
Process
- If argument provided, use it as the learning content
- If no argument, ask: "What should we capture?"
- Determine the appropriate entry type:
- decision: Architectural or design choice made
- pattern: How we do things here (reusable approach)
- gotcha: Trap to avoid, something that caught us
- context: Background info, not actionable
- Extract or ask for relevant topics (tags for categorization)
- Use the
mcp__megg__learntool to save
Example
User: /megg-learn always use absolute paths in hooks
→ Capture as pattern with topics like hooks, paths
Important
- Keep entries concise and actionable
- Topics should be 1-3 words each
- Title should be short (5-10 words max)
- Content can include examples or context