claude-code-help

claude-code-help

Comprehensive help skill for Claude Code. 40 playbooks, 638 indexed items, 4-level progressive disclosure, all citation-backed.

Quick Start

npx claude-code-help init

Then open Claude Code and type /claude-code-help.

What You Get

Content	Count	Description
Playbooks	40	Step-by-step guides across 7 categories
Progressive levels	5	L0 quickstart through L4 ops
Reference files	3	CLI commands, settings, permissions
Citation sources	55	Links to official documentation
Indexed items	638	Commands, flags, shortcuts, settings, env vars

Installation

Project-local (recommended)

Installs to .claude/skills/ in your current project:

npx claude-code-help init

Global

Installs to ~/.claude/skills/ so the skill is available in all projects:

npx claude-code-help init --global

Overwrite existing

If you already have the skill installed and want to update:

npx claude-code-help init --force

Usage Examples

Once installed, use /claude-code-help in Claude Code. Ask questions like:

• "How do I configure MCP servers?" -- routes to the MCP setup playbook
• "Show me all keyboard shortcuts" -- loads the L1 core reference with shortcut tables
• "Help me set up git hooks" -- opens the hooks configuration playbook
• "What permission modes are available?" -- pulls from the permissions reference

The skill uses a routing brain (SKILL.md) to match your question to the right content level and load only what is needed, keeping context usage minimal.

CLI Reference

claude-code-help init [options]

Flag	Short	Description
--global	-g	Install to ~/.claude/skills/ (global)
--force	-f	Overwrite existing installation
--help	-h	Show help message
--version	-v	Show version number

Contents Overview

Progressive Levels

• L0 -- Quickstart (first 5 minutes)
• L1 -- Core commands, shortcuts, configuration
• L2 -- Power features, all flags, MCP, hooks
• L3 -- Advanced workflows, multi-agent, custom slash commands
• L4 -- Ops, CI/CD, monitoring, security hardening

Playbook Categories

• Getting Started (5 playbooks) -- installation, first project, configuration
• Core Workflows (7 playbooks) -- editing, git, testing, debugging
• Configuration (6 playbooks) -- settings, CLAUDE.md, permissions, models
• MCP & Tools (5 playbooks) -- server setup, custom tools, browser automation
• Hooks & Automation (5 playbooks) -- git hooks, CI integration, custom scripts
• Advanced Patterns (7 playbooks) -- multi-agent, memory, large codebases
• Troubleshooting (5 playbooks) -- common errors, performance, recovery

Reference Files

• CLI Command Reference -- 9 commands, 48 flags, 8 MCP subcommands
• Settings Reference -- 55+ settings keys, 117 environment variables
• Permissions Reference -- 6 permission modes, tool allowlists

How It Works

The skill uses fork context loading: SKILL.md is always loaded (308 lines), but the bulk of the content (progressive levels, playbooks, references) is loaded on-demand only when relevant to your query. This keeps the context footprint small while providing comprehensive coverage.

SKILL.md (always loaded, ~12KB)
  |
  +-- routes to --> progressive/L0-L4 (on-demand)
  +-- routes to --> playbooks/*.md (on-demand)
  +-- routes to --> reference/*.md (on-demand)

Development

Sync skill files from source

npm run sync-skill

Run tests

npm test

Verify before publishing

node scripts/verify-skill.js

Publish

git tag v0.1.0
git push origin v0.1.0
# CI handles npm publish with provenance

License

MIT