What problem does it solve? Answering structural questions about a codebase—what depends on X, how A connects to B, what a subsystem does—by grepping and reading raw files burns context and time. This Skill routes structural questions to a code knowledge graph first and reserves file reads for the files that actually matter. ## Core Features & Use Cases - Graph-first querying: Detects and queries a codebase knowledge graph (e.g. graphify via CLI or MCP tools like query_graph, shortest_path, get_node) for impact analysis, path queries, and concept explanation. - Deterministic fallback: Builds a cheap structural repo map from git ls-files, symbol listings, and import edges when no graph exists. - Workflow integration: Grounds /spec decomposition, /feature impact analysis, and /fix debugging in real dependency data while preserving Read Before You Write discipline. - Use Case: Before editing a shared module, ask the graph what depends on it, verify INFERRED edges by reading the flagged callers, then make the change with full impact awareness. ## Quick Start Ask the agent to use the code-query skill to find what depends on the module you are about to change before editing it.
