What problem does it solve?
Long-running agent tasks lose all context when a conversation ends, forcing you to rebuild state, memory, and progress from scratch. This Skill snapshots agent sessions so work can be paused and resumed later with full context intact.
Core Features & Use Cases
- Session Snapshots: Save the complete agent state, including memory entries, learning trajectories, metrics, and task progress, via the session_save tool.
- Session Restoration: List, inspect, and restore previous sessions by ID to resume work exactly where it stopped.
- Lifecycle Hooks: Automatically save state at conversation end and restore context at conversation start using session-start, session-end, and session-restore hooks.
- Use Case: During a multi-day refactoring project, checkpoint progress each evening with a session save, then restore the snapshot the next morning to continue with all todos, memory, and metrics preserved.
Quick Start
Ask the agent to save the current session state, then later request it to list saved sessions and restore the one you want to resume.