Codebase Graph is a skill that analyzes unfamiliar repositories and produces structured visual and textual representations of their architecture. It parses a project’s directory structure, import statements, and configuration files such as package.json using AST-based analysis to build a dependency model of the codebase.
From that model, the skill generates a standalone, interactive D3.js force-directed graph that shows how modules relate to one another. It also produces detailed Markdown summaries covering entry points, refactor suggestions, and module relationships. Beyond a basic file-tree view, it detects circular dependencies, identifies likely dead code, and ranks modules by their centrality to the overall architecture, helping developers understand where execution begins and where technical debt concentrates.
Supported languages include TypeScript, JavaScript, and Python. The skill is aimed at situations where a developer needs to understand a large, unfamiliar codebase without manually tracing files — for example, during onboarding or when planning a major refactor. Because the skill has no listed environment variables or external service dependencies, it operates against the local repository files it is given access to.
This is not a runtime monitoring or database introspection tool. It does not connect to Postgres or any other database, so it is the wrong choice if the goal is querying live data rather than understanding source code structure.