brain-health

Knowledge graph health is measured by applying rubric-driven metrics across themes, density, hubs, co-occurrence, dedup pressure, and entity landscape. This generates a prioritized findings list with severity levels and suggested actions for maintenance.

Detecting knowledge graph drift involves analyzing a 7-day window to produce per-theme deltas, drift patterns, and cross-metric patterns. This enables longitudinal tracking and cross-run comparisons suitable for executive review.

Knowledge graph reporting can be automated to generate rubric-scored health reports and persist the full output to a timestamped markdown file. This supports auditing and stakeholder sharing.

Stale queues and dedup pressure are reviewed using automated tool integrations like review_stale and dedup_review. These assess graph health, surface maintenance needs, and generate actionable insights for optimization.

Knowledge graph health analysis operates with no external dependencies or components. It relies on built-in automated tool integrations like thought_stats, list_thoughts, list_entities, and serendipity_digest to assess the graph directly.

Rubric-driven metrics for knowledge graph maintenance should be used during regular health reviews, drift detection, hub stability checks, and pattern discovery. They guide optimization by surfacing theme attention, synthesis health, and entity landscape status.