codebase-scan

Generate .claude/CLAUDE.md maps and per-module rules from codebases.

Updated Apr 22, 2026

Search Tags:#automation #documentation #codebase #brain #context-loading #module-mapping #nx-monorepo

Ready to run

npx skills add https://github.com/Shockwave2k/team-claude-skills --skill codebase-scan-shockwave2k

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: codebase-scan
Source: https://github.com/Shockwave2k/team-claude-skills/tree/main/skills/shared/codebase-scan
Command: npx skills add https://github.com/Shockwave2k/team-claude-skills --skill codebase-scan-shockwave2k

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates building a structured, per-project brain for codebases by generating a high-level CLAUDE map and per-module rules, enabling fast context loading and targeted knowledge reuse.

Core Features & Use Cases

Generate a central .claude/CLAUDE.md high-level map that describes the project architecture.
Create per-module rules at .claude/rules/<module>.md so Claude loads module-specific context on demand.
Works best with NX monorepos and service-oriented repos; run once per project and re-run after major restructures.

Quick Start

Run codebase-scan on your repository to generate the initial brain files.

Frequently Asked Questions about codebase-scan

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

FAQPage Schema

How do I generate a per-project brain for an NX monorepo codebase?▼

To generate a per-project brain, scan your NX monorepo codebase to automatically produce a high-level CLAUDE map and per-module rules, enabling fast context loading and targeted knowledge reuse.

How does on-demand context loading work for service-oriented repositories?▼

On-demand context loading works by generating per-module rules at .claude/rules/<module>.md, so Claude loads module-specific context only when needed, reducing overhead when navigating service-oriented repositories.

Can I use codebase scanning to map module boundaries in a service-oriented repo?▼

Yes, you can use codebase scanning to analyze repository structure, entry points, and module boundaries in service-oriented repos, outputting read-only brain files under .claude/ for structured architecture mapping.

What is the best way to automate CLAUDE.md generation for a large codebase?▼

The best way to automate CLAUDE.md generation is to run a codebase scan that analyzes repository structure and entry points, outputting a central .claude/CLAUDE.md high-level map describing the project architecture.

Do I need to re-run codebase scanning after restructuring my monorepo?▼

Yes, you should re-run codebase scanning after major restructures to keep brain files accurate, as the generated .claude/CLAUDE.md map and per-module rules must reflect updated module boundaries and architecture.

What are the limitations of automated module mapping for non-NX repositories?▼

Automated module mapping works best with NX monorepos and service-oriented repositories, so non-NX or non-service-oriented codebases may not benefit fully from the per-module rules and context loading generation.