cc-onboard

Scan existing repositories to generate CLAUDE.md, .mcp.json, and rules.

Updated Mar 17, 2026

Search Tags:#automation #onboarding #codebase #claude-code #architecture-detection #knowledge-store #rules-generation

Ready to run

npx skills add https://github.com/SPRAGE/dev-template --skill cc-onboard

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: cc-onboard
Source: https://github.com/SPRAGE/dev-template/tree/main/templates/python/.claude/skills/cc-onboard
Command: npx skills add https://github.com/SPRAGE/dev-template --skill cc-onboard

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Repositories often have code but lack Claude Code configuration, causing slow onboarding and inconsistent rule generation. This skill automates the end-to-end setup by scanning the codebase bottom-up to generate CLAUDE.md, .mcp.json, and .claude/rules/, and by populating the knowledge store with architecture and convention data detected from the code.

Core Features & Use Cases

Scans the repository to infer languages, frameworks, and conventions.
Generates CLAUDE.md, .mcp.json, and a rules set tailored to the project.
Populates the knowledge store with detected architecture and coding conventions for onboarding new contributors.

Quick Start

Initiate onboarding on a repository with Claude Code not configured and run the onboard workflow to generate CLAUDE.md, .mcp.json, and knowledge rules.

Frequently Asked Questions about cc-onboard

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

FAQPage Schema

How do I generate Claude Code configuration files for an existing repository?▼

To generate Claude Code configuration for an existing repository, this skill scans the codebase bottom-up to automatically create CLAUDE.md, .mcp.json, and .claude/rules/ files tailored to your project's architecture.

What is the best way to automate Claude Code onboarding for a codebase with no prior setup?▼

Automating Claude Code onboarding involves scanning the existing code to infer languages, frameworks, and conventions, then populating a knowledge store and generating exact configuration structures like CLAUDE.md and .mcp.json.

Can I infer coding conventions and architecture automatically when setting up Claude Code?▼

Yes, you can infer coding conventions and architecture automatically by using four agent findings that scan the codebase bottom-up to detect language, framework, structure, and commands, then bootstrap rules from that data.

Does Claude Code onboarding work on repositories with existing code but no configuration?▼

Yes, Claude Code onboarding specifically targets repositories with existing code but no Claude Code configuration, inferring architecture and conventions directly from the code to bootstrap necessary rules.

What specific files are generated when bootstrapping Claude Code into a project?▼

Bootstrapping Claude Code generates a CLAUDE.md file, a .mcp.json configuration, a .claude/rules/ directory, and a populated knowledge store containing detected architecture and convention data.

How does environment-variable handling work during Claude Code onboarding?▼

During Claude Code onboarding, environment-variable handling respects safe defaults, ensuring that the generated CLAUDE.md and .mcp.json files do not expose sensitive configuration inferred from the codebase.