action-analyze-codebase

To analyze a codebase for architecture and dependencies, you can use a structured analysis tool that reads core documentation and manifests to produce a concise repository overview and detailed cross-module relationship artifacts.

Generating structured documentation artifacts for a large repository is best handled by deep partitioned analysis with subagent delegation, which writes structured outputs detailing entrypoints, tooling, and cross-module relationships to a designated directory.

Yes, you can get a quick codebase summary using surface analysis, which rapidly reads core documentation and manifests to return a concise repository overview without performing a deep partitioned scan.

You extract cross-module relationships and entrypoints by running a deep partitioned analysis, which maps out the architecture, dependencies, and tooling to produce structured artifacts for other skills to consume.

Yes, this codebase analysis approach works with repositories of any size by supporting surface analysis for quick summaries and deep partitioned analysis with subagent delegation for larger, more complex codebases.

The structured analysis artifacts are saved under the ./analysis/ directory, providing a standardized output location that other skills can consume to generate documentation or answer questions.