/kb-memory-commit

Commit to long-term memory: $ARGUMENTS <ultrathink> Memory transforms experience into wisdom. What we remember shapes what we can imagine and achieve. </ultrathink> <megaexpertise type="memory-architect"> The assistant should use the memory-architect agent to capture and store context in ContextFrame with proper metadata, embeddings, and relationships. </megaexpertise> <context> Context description: $ARGUMENTS Current directory: !pwd Recent git commits: !git log --oneline -5 2>/dev/null || echo "No git history" Memory store path: .claude/memory/ Commit types: implementation, decision, incident, learning, milestone </context> <requirements>

• Capture comprehensive context summary
• Generate semantic embeddings if available
• Link to related knowledge base entries
• Include git context (commits, branch, files)
• Tag appropriately for future retrieval
• Store in ContextFrame with proper metadata </requirements>

<actions> 1. Analyze Current Context: - Recent git activity - Changed files - Related knowledge entries - Session activities

2. Determine Memory Type:

   • implementation: Feature or code implementation
   • decision: Architectural or design decision
   • incident: Bug, issue, or incident resolution
   • learning: New knowledge or pattern discovered
   • milestone: Project milestone or major achievement

3. Launch Memory Architect Agent:

   Use the Task tool with subagent_type="general-purpose" to invoke the memory-architect agent.
   
   Prompt:
   "Commit the following context to long-term memory:
   Description: $ARGUMENTS
   
   Please:
   1. Analyze recent git commits and changes
   2. Identify the memory type (implementation, decision, incident, learning, or milestone)
   3. Create a comprehensive context summary (2-5 paragraphs)
   4. Extract relevant tags, components, and metadata
   5. Search for related knowledge base entries to link
   6. Create a ContextFrame record with:
      - UUID
      - Content summary
      - Metadata (type, tags, component, git info)
      - Relationships to knowledge entries
      - Status: active
   7. Generate semantic embedding if embedding service is configured
   8. Store in .claude/memory/contextframe.lance
   9. Confirm storage and provide UUID for future reference
   
   If ContextFrame is not installed, provide instructions for installation."
   

4. Present Results:

   • Memory UUID
   • Commit type and tags
   • Embedding status
   • Linked knowledge entries
   • Storage confirmation </actions>

The assistant should make memory commits comprehensive yet efficient, ensuring critical context is preserved with rich metadata for future semantic retrieval.