What problem does it solve? Maintaining a repository knowledge graph manually is error-prone: dangling edges, orphan nodes, and unclustered nodes accumulate silently, and answering questions like "what uses this module" requires tedious manual searching. This Skill automates health checks, gap analysis, and natural-language graph queries. ## Core Features & Use Cases - Health Monitoring: Runs health checks and gap analysis on the merged graph layer, reporting dangling edges, duplicate node IDs, stub nodes, orphan nodes, and unclustered nodes in a formatted summary. - Interactive Repair Workflow: Presents actionable options (auto-fix dangling edges, discover orphan connections, assign clusters, full rebuild) and loops until issues are resolved or skipped. - Natural Language Graph Queries: Answers questions like "show me the call flow for X", "what depends on X", or "what skills load for this intent" by routing to the appropriate query script. - Use Case: After refactoring skills or agents in a repository, ask to "check the graph" to detect broken edges, then auto-fix them and verify the graph returns to a healthy state. ## Quick Start Ask the agent to check the knowledge graph health and show any dangling edges or orphan nodes that need attention.
