chat-compactor

To create a session handoff summary, compact your chat history into structured markdown documenting decisions, outcomes, and next steps. This preserves context so the next AI agent can continue coding or debugging without re-explaining the background.

Context compaction summarizes a long chat history into a machine-readable format. You need it when the context window gets long during a debugging sprint, ensuring continuity and preventing lost decisions before ending an AI session.

Document failed approaches by generating a session summary that explicitly records dead-ends and blockers. This prevents the next AI agent from re-deriving context or retrying methods that already failed, saving debugging cycles.

Yes, you can save a structured markdown handoff file directly to your project. The summary is automatically saved to a .handoff/session-[topic]-[date].md file, creating a scannable document for later agent retrieval.

The best way to prepare an actionable checklist is to compact the session into a prioritized list of next steps. This ordered checklist guides immediate follow-through after a task switch, capturing what worked and what should happen next.

Yes, session summarization works without additional dependencies. The compaction process relies solely on your existing AI agent to process the chat history and output the structured markdown file directly into your project directory.