trw-memory-optimize

You can prune stale TRW learnings by auditing the memory store to identify obsolete entries, then interactively confirming a plan that marks deletions as obsolete and recalibrates impact scores to reduce noise.

Consolidating duplicate memory entries and rebalancing tags involves detecting redundant learnings, creating compendium entries for them, and normalizing inconsistent tags across the learning store after explicit user confirmation.

Interactive memory optimization prevents data loss by building an optimization plan first, requiring explicit user confirmation before executing changes, and safely marking deletions as obsolete rather than permanently removing them.

Yes, you can synchronize documentation and delivery state after modifying learnings; the optimization workflow automatically syncs AGENTS.md and delivery traces to preserve accurate workflow records post-change.

Auditing inconsistent tagging and duplicates in a learning store uses TRW recall and index data to identify stale entries, detect redundant learnings, and surface tag inconsistencies for normalization.

Before and after reports reveal active entries, obsoleted items, consolidations, and tag normalization outcomes, providing a clear comparison of the memory store's state before and after the optimization changes.