codebase-graph

A dependency graph visualizes relationships between files, modules, functions, and classes using an ontology-based approach. This Skill automates graph generation by analyzing your codebase with AST parsing, creating nodes for code entities and edges for imports, exports, calls, extends, implements, and contains relationships. The graph is generated automatically on project start and updated incrementally as files change.

The graph represents seven edge types: imports, exports, calls, extends, implements, uses, and contains. Node types include files, modules, functions, classes, types, and variables. This ontology-based schema enables you to trace import/export mappings, function call chains, inheritance hierarchies, and module dependencies across your entire codebase.

Yes. Once the graph is constructed, it provides a queryable interface for path discovery and dependency tracing. You can explore how modules depend on each other, discover function relationships, and understand architecture insights by traversing nodes and edges in the knowledge graph.

The Skill uses token-efficient graph construction with caching and versioning to handle large systems. It's built on ts-morph for TypeScript AST analysis and processes incremental file changes rather than full rebuilds, keeping analysis fast as your codebase grows.

Use a codebase graph when you need to visualize and query complex relationships across modules, understand architecture at scale, trace multi-step dependencies, or perform impact analysis. Unlike flat static analysis, the ontology-based graph structure supports sophisticated queries and reveals patterns hidden in unstructured code relationships.

Yes. The implementation uses ts-morph for TypeScript AST parsing, making it native to TypeScript projects. It automatically triggers on project start and maintains an updated graph as TypeScript files change.