What problem does it solve? Knowledge scattered across local files, chat sessions, GitHub, Linear, and databases becomes hard to find and easy to duplicate. This Skill provides a structured workflow for ingesting, deduplicating, syncing, and retrieving knowledge across multiple storage layers. ## Core Features & Use Cases - Multi-Layer Knowledge Architecture: Routes information to the right store, from Claude Code memory files and the MCP memory knowledge graph to GitHub/Linear, a knowledge base repo, and external stores like Supabase. - Ingestion Workflow: Classifies new knowledge by type, deduplicates against existing entries, stores it in the appropriate layer, and updates indexes. - Sync Operations: Syncs conversation sessions, workspace state, and cross-source exports (Claude/ChatGPT/Grok conversations, bookmarks, GitHub activity) into a durable knowledge base repo. - Use Case: After a long planning session, ask the assistant to save the key decisions to the knowledge base; it will classify each fact, check for duplicates, write memory files and MCP memory entities, and commit a session summary to the KB repo. ## Quick Start Save the key decisions from this conversation to my knowledge base and sync it with my existing memory.
