trw-memory-audit

Auditing self-learning memory health requires analyzing the memory layer to compute metrics on tag distribution, impact, staleness, and duplicates. This surfaces quality issues and actionable insights for pruning and consolidating learnings.

Memory tag distribution analysis identifies coverage gaps and orphan tags across learnings. You need it during sprint cycles to inform consolidation, improve memory quality, and ensure your self-learning system maintains comprehensive coverage.

Yes, you can generate top recommendations to prune, consolidate, and rename tags, as well as fill coverage gaps. The audit reads the index and related data, computes health metrics, and outputs a structured report with actionable steps.

No, the memory audit enforces strict read-only behavior to avoid modifying learnings. It reads the index and related data to compute metrics and output a structured health report without altering the underlying memory layer.

Checking assertion health and observability involves running observability and assertion health checks to ensure transparent memory health reporting. This validates the integrity of your learnings and surfaces any underlying quality issues.

The best way to surface actionable insights is computing impact, staleness, and duplicates metrics across the memory layer. This structured analysis informs targeted pruning and consolidation strategies to systematically improve memory quality.