cc-init

Bootstrap a lean Claude Code configuration with CLAUDE.md, settings, and hooks.

1|1|Updated Apr 2, 2026
One-click install
npx skills add https://github.com/MichaelvanLaar/claude-code-config-skills --skill cc-init
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cc-init
Source: https://github.com/MichaelvanLaar/claude-code-config-skills/tree/main/.claude/skills/cc-init
Command: npx skills add https://github.com/MichaelvanLaar/claude-code-config-skills --skill cc-init

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Creates a minimal, safe Claude Code configuration for a new or unconfigured project so repositories start with a compact CLAUDE.md, hardened permissions, and sensible hooks instead of verbose, token-heavy, and error-prone defaults.

Core Features & Use Cases

  • Creates a .claude/settings.json with deny rules and cost-optimization env defaults to protect secrets and reduce token usage.
  • Generates a concise, 20–40 line CLAUDE.md tailored to the repo or interactive answers, with placeholders for unknown commands.
  • Optionally creates AGENTS.md, updates .gitignore, and wires a sync script plus pre-commit hook to keep a "Key Config Files" table in sync.
  • Scans the repo for language/tooling clues (package managers, linters, formatters, CI) and only adds PostToolUse hooks when a formatter is confidently detected.
  • Use case: initialize a brand-new project or add a safe baseline to a minimal repo before writing code or setting up CI.

Quick Start

Run the skill by typing /cc-init and supply a brief one-line project description or follow the interactive prompts.

Frequently Asked Questions about cc-init

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

FAQPage Schema
How do I bootstrap a Claude Code configuration for a new project?

To bootstrap a Claude Code configuration, you can run a skill that generates a compact CLAUDE.md, hardened .claude/settings.json with deny rules, and cost-optimization environment defaults. It scans your repo for tooling clues to safely wire basic hooks without destructive operations.

What's the best way to set up safe Claude Code permissions before writing code?

Setting up safe Claude Code permissions involves generating a .claude/settings.json file with deny rules and cost-optimization environment defaults. This protects secrets and reduces token usage by establishing hardened boundaries before any code is written or CI is configured.

How does Claude Code detect project tooling to configure hooks automatically?

Claude Code hook configuration works by scanning the repository for language and tooling clues like package managers, linters, and formatters. It only adds PostToolUse hooks when a formatter is confidently detected, avoiding error-prone defaults in the generated settings.

Do I need an empty repository to initialize a minimal Claude Code config?

You do not need a completely empty repository to initialize a minimal Claude Code config. This process applies to both empty and minimal repositories, adding a concise 20–40 line CLAUDE.md, safe permissions, and basic hooks without overwriting existing code.

Can I generate a CLAUDE.md file with placeholders for unknown commands?

You can generate a concise CLAUDE.md file tailored to your repository or interactive answers. The resulting 20–40 line file includes placeholders for unknown commands, ensuring a compact, token-light configuration instead of verbose defaults.

What files does a Claude Code bootstrap process create in an unconfigured repo?

A Claude Code bootstrap process creates a .claude/settings.json file with deny rules, a short CLAUDE.md, and an optional AGENTS.md. It also updates .gitignore and can wire a sync script plus pre-commit hook to keep a Key Config Files table in sync.