context-and-continuity

What problem does it solve?

Long-running agents lose valuable intermediate discoveries when context windows shrink or sessions restart, causing rework and missed insights. This Skill provides explicit checkpointing and memory configuration patterns so agents can save, retrieve, and review important findings despite context pressure or restarts.

Core Features & Use Cases

• Explicit checkpoint API: Save named findings, retrieve by key, or list all checkpoints to ensure important discoveries persist beyond the active message window.
• Auto-checkpointing and windowing guidance: Built-in thresholds and recommended iteration budgets help avoid context compaction and surface when to persist observations.
• Cross-session persistence: Integrate enhanced memory tiers with a database path to survive agent restarts and support semantic recall for multi-session workflows.
• Use Case: Run a multi-step research agent that discovers API endpoints and configuration details across dozens of iterations and persist each major finding so the final report compiles all confirmed results reliably.

Quick Start

Start an agent configured with the checkpoint tool enabled and enhanced persistent memory so it saves key discoveries across a long-running task.