What problem does it solve? It prevents loss of context between long AI-assisted work sessions by durably recording decisions, completed work, and gotchas so the next session resumes without re-litigating prior choices. ## Core Features & Use Cases - Narrative Session Logging: Appends a structured checkpoint block (goal, completed items, decisions, files touched, gotchas) to .memory/SESSION-LOG.md. - TODO Synchronization: Updates TODOS.md in-place, removing completed items and appending new tagged open items, supporting both flat and .work/plan/ formats. - Knowledge Promotion & Triage Refresh: Proposes eligible facts for local or global KNOWLEDGE.md and runs idempotent cache, log-rotation, and triage-pipeline commands. - Use Case: At the end of a multi-hour coding session, run /checkpoint to capture why decisions were made, then /clear and paste the printed re-entry prompt to start the next session with full context. ## Quick Start Ask the AI to run /checkpoint at the end of your work session to write the session log, sync TODOS.md, and print a re-entry prompt.