What problem does it solve? Persistent memory systems degrade silently — files go stale, routing strategies fall back without notice, and extraction jobs fail in the background. This Skill runs a structured audit so you can see the actual state of your memory files, plugin infrastructure, and active configuration in one report instead of discovering problems after bad recall. ## Core Features & Use Cases - Memory File Audit: Scans the entire memory directory, reports fresh vs. stale counts, and flags missing starter files (me.md, technical-pref.md, preferences.md) with size and age details. - Routing Quality Diagnosis: Reports configured vs. effective routing strategy, live aggregate metrics (cap saturation, score spread), and last evaluation results with honest caveats about token_overlap limitations. - Infrastructure Checks: Validates directory existence, diary file counts, in-flight and failed extractions, unprocessed learnings, and last dream consolidation date. - Use Case: After noticing your assistant's recall feels off, run the health audit to discover that 8 of 27 memory files are stale and three extraction jobs failed — then follow the generated recommendations to run dream consolidation. ## Quick Start Ask the assistant to run a health check on my memory plugin and show me which files are stale or missing.
