What problem does it solve? When working across multiple AI sessions, context about past decisions, prior measurements, and unfinished work gets lost. This Skill retrieves stored session history so you can answer historical questions and resume work without guessing or redoing completed analysis. ## Core Features & Use Cases - Semantic History Search: Query stored session history semantically rather than with regex, scoped by project and time window, with batched independent questions. - Grounded Interpretation: Distinguish historical evidence from current repository state, avoiding the mistake of treating past decisions or measurements as proof of today's state. - Cursor-Based Paging: Page through recent sessions using returned cursors and follow up with returned IDs instead of invented ones. - Use Case: You return to a project after a week and need to know why a specific architecture decision was made and what tasks were left incomplete. Use this Skill to search prior sessions, retrieve the relevant decision records, and resume the unfinished work with full context. ## Quick Start Ask the assistant to recall what was decided and left unfinished in this project during last week's sessions.
