chat-compactor

To preserve AI agent context for a session handoff, generate a structured Markdown document capturing decisions, current state, and next steps. This allows the next agent to resume work without requiring prior context re-explanation.

The best way to summarize a coding or debugging session is to produce a machine-readable handoff file with distinct sections for context, decisions, current state, and next steps, ensuring continuity for future work.

You save a session summary by running the compactor to generate a Markdown file named session-[topic]-[date].md in the project root, allowing the next AI agent to resume debugging directly from the saved state.

Yes, you can use a structured session summary for ongoing research continuity. It captures rationale and blockers in a consistent format, reducing retry time and re-explanation when switching tasks or ending sessions.

The session handoff document uses a Markdown format that organizes decisions, rationale, blockers, context, current state, and next steps into distinct sections for machine-readable preservation.

Generating an AI handoff summary requires no external dependencies. The skill operates independently to produce a Markdown document in the project root without needing additional components or libraries.