note_taker
SKILL.md
Note Taker Skill
Purpose
The Scribe's Ledger enables the agent to maintain persistent memory across sessions by:
- Capturing - Recording session trajectories, decisions, and outcomes
- Summarizing - Extracting key insights from execution traces
- Retrieving - Loading relevant past knowledge for new tasks
- Pattern Mining - Identifying reusable solutions and anti-patterns
System Prompt Additions
When this skill is active, add these guidelines to the LLM context:
# Note Taker Guidelines
When working with knowledge management:
- Use `note_taker.summarize_session` to capture completed work before switching contexts
- Use `note_taker.update_knowledge_base` to save discovered patterns and solutions
- Use `note_taker.search_notes` to recall past approaches before reinventing
- Always record:
- Key decision points and rationale
- Failed approaches and why they failed
- Final working solution and its context
- Files modified and their purpose
Commands
| Command | Description |
|---|---|
note_taker.summarize_session |
Summarize current session into markdown notes |
note_taker.update_knowledge_base |
Save extracted knowledge for future retrieval |
note_taker.search_notes |
Search existing notes and knowledge |
note_taker.extract_patterns |
Identify reusable patterns from sessions |
Key Directories
| Directory | Purpose |
|---|---|
.data/knowledge/sessions/ |
Session summaries (auto-generated, git-ignored) |
assets/knowledge/harvested/ |
Extracted reusable knowledge |
assets/knowledge/patterns/ |
Identified patterns and anti-patterns |
Related Documentation
- CCA Paper - Cognitive Scaffolding Architecture
- Skills Documentation - Comprehensive skill guide