codebase-skeleton

Compress source code repositories into compact skeletons with imports and signatures.

Updated Apr 19, 2026

Search Tags:#context #onboarding #codebase #code-review #signatures #ast #skeleton

Ready to run

npx skills add https://github.com/crycriM/hermes-skills --skill codebase-skeleton

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: codebase-skeleton
Source: https://github.com/crycriM/hermes-skills/tree/main/software-development/codebase-skeleton
Command: npx skills add https://github.com/crycriM/hermes-skills --skill codebase-skeleton

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Compress a source code repository into a compact skeleton that captures structure and signatures without function bodies, enabling small-context models to understand codebases efficiently.

Core Features & Use Cases

Extract imports, class and function signatures with type hints, decorators, and truncated docstrings.
Produce a compact "skeleton" tree or JSON-like representation for onboarding, code review, and feature planning.
Handle syntax errors gracefully and skip unparsable files to maintain a usable outline.

Quick Start

Run the codebase-skeleton script on your repository to generate a compact signature map of the codebase.

Frequently Asked Questions about codebase-skeleton

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I extract function signatures and class structure from a Python codebase for LLM context?▼

You can extract function signatures and class structure by compressing a source code repository into a compact skeleton that captures imports, decorators, and signatures without function bodies. This produces a structured outline suitable for small-context LLMs to understand the codebase efficiently.

What is a codebase skeleton and when do I need one for code review?▼

A codebase skeleton is a compact representation of a repository that captures structure and signatures without function bodies. You need one for code review, onboarding, or architecture planning when you want to understand the overall code layout without loading full implementations.

Can I generate a compact codebase skeleton for multi-language repositories?▼

Yes, you can generate compact skeletons across Python and multi-language repos. The extraction process pulls imports, classes, functions, decorators, and truncated docstrings into a reproducible tree or JSON-like format regardless of the language mix.

Does the skeleton extraction process handle syntax errors in source files?▼

Yes, the skeleton extraction process handles syntax errors gracefully by skipping unparsable files to maintain a usable outline. This ensures that broken files in a repository do not halt the generation of the overall signature map.

What is the best way to onboard a small-context LLM onto a large existing codebase?▼

The best way is to run a skeleton extraction script on the repository to generate a compact signature map. By extracting type hints, imports, and truncated docstrings without bodies, the resulting structured output fits within small-context model limits.