What problem does it solve? Navigating an unfamiliar codebase to answer architecture or dependency questions is slow when you must read files one by one. This Skill answers questions about the codebase by querying a pre-built knowledge graph, so you get targeted answers grounded in actual files, functions, and relationships without dumping the whole project into context. ## Core Features & Use Cases - Graph-Based Q&A: Searches node names, summaries, and tags in the knowledge graph JSON to find code entities matching your query, then follows edges (imports, calls, depends_on) to build a 1-hop subgraph around them. - Freshness Validation: Compares the graph's recorded git commit against HEAD and checks committed and working-tree diffs, warning you when the graph may be stale and suggesting a refresh. - Layer-Aware Answers: Uses architectural layer metadata to explain which parts of the system are relevant and why, linking concepts to concrete file paths. - Use Case: Ask "which modules depend on the billing service?" and get an answer traced through import and call edges, with exact file locations, instead of grepping the entire repository manually. ## Quick Start Ask a question about the codebase, such as "how does the authentication flow connect to the API layer?", and the Skill will query the knowledge graph and answer with specific files and relationships.
