superclaw-autodream

OpenClaw memory consolidation and de-duplication can be automated by triggering a four-phase process that merges, repairs, and indexes topic files. This organizes memory sources to keep context fresh and consistent.

Memory consolidation handles incremental updates by gathering signals from recent daily notes and completed tasks over the past 7 days. It then merges these new signals into existing topic files while de-duplicating redundant entries.

Yes, memory indexing is designed to be safe and idempotent, meaning you can run it repeatedly without duplicating results. It tracks state updates in a dedicated JSON file and maintains a run history to prevent redundant processing.

Automated memory management requires existing memory files in your MEMORY.md file and corresponding topic files. Once these sources are present, the consolidation process can orient, gather signals, and prune automatically.

The best way to trigger automatic memory consolidation is via an agent-check or manual invocation. Once started, the system automatically processes the four phases—orient, gather signal, consolidate, prune & index—to update your memory map.

The prune and index phase is needed to remove outdated or redundant memory sources and rebuild a compact, searchable index. This ensures the memory map remains up-to-date and consistent for future context retrieval.