awesome-claude-code-setup

👋 What is AwesomeClaude?

Claude Code is already incredible — but out of the box, it doesn't know anything about your project. Every new session, you're starting from scratch. AwesomeClaude fixes that.

One install gives Claude an instant understanding of your codebase, a library of ready-to-use workflows, and smart automations that run quietly in the background. It's built on Anthropic's own Claude Code Best Practices and installs in under a minute.

• ✅ Instant project understanding — One command (chp) gives Claude everything it needs to know about your project. No explaining, no re-reading files every session.
• ✅ Complete workflows as slash commands — /start-feature, /debug-issue, /pre-deploy-check — whole multi-step workflows in a single command.
• ✅ Catches problems early — Automatically spots security issues, overly complex code, and common bug patterns before they ship.
• ✅ Runs quietly in the background — Auto-formats your files after every edit, notifies you when a long task finishes. Set it once, forget about it.
• ✅ Connects to your tools — Browser automation, live docs, GitHub, and more — available when you need them, out of the way when you don't.

🚀 Quick Install (30 seconds)

# One-line install
curl -sSL https://raw.githubusercontent.com/cassler/awesome-claude-code-setup/main/setup.sh | bash && source ~/.zshrc

# Or if you prefer to look before you leap:
git clone https://github.com/cassler/awesome-claude-code-setup.git && cd awesome-claude-code-setup && ./setup.sh && source ~/.zshrc

That's it! The setup script handles everything — it detects what you already have, installs only what's needed, and works across all your projects from day one.

📦 What Gets Installed?

1. Helper scripts → ~/.claude/scripts/ (the engine under the hood)
2. Slash commands → ~/.claude/commands/ (the workflows you'll actually type)
3. Shell aliases → Added to your .zshrc or .bashrc (short commands like chp)
4. Global config → ~/.claude/CLAUDE.md (tells Claude about the tools available)
5. Hooks config → ~/.claude/settings.json (auto-format + notifications, skipped if you already have one)
6. Scoped rules → ~/.claude/rules/ (language-specific guidance that loads automatically)

---

🤖 Which Claude Model Should I Use?

Model	Best For
Opus 4.7	Big, complex tasks — large refactors, tricky bugs, deep reasoning
Sonnet 4.6	Your everyday go-to — fast, smart, handles most things great
Haiku 4.5	Quick questions, simple edits, when speed matters most

Switch models anytime with /model in Claude Code, or use the model picker in the desktop app. Sonnet 4.6 is the default and a great starting point.

---

📝 Slash Commands — Your Workflow Superpowers

Just type / in Claude Code to pull up any of these. They're full workflows — not just prompts — that walk Claude through each step automatically.

Building Features

• ✨ /start-feature — Creates your GitHub issue, branch, and draft PR in one shot
• 🐛 /debug-issue — Traces a bug step by step with root cause analysis
• ✅ /pre-review-check — Makes sure your code is ready before asking for a review
• 🚢 /pre-deploy-check — Runs through a full production-readiness checklist
• 🔁 /autofix-pr — Reads CI failures and review feedback, fixes what it can, pushes

Understanding Your Code

• 🧠 /understand-codebase — Get up to speed on any project quickly
• 🔍 /explore-module — Deep dive into how a specific module works
• 📦 /analyze-dependencies — See what your project depends on and flag anything risky
• 🌐 /api-documenter — Auto-generate documentation for your API endpoints
• 🔧 /refactor-assistant — Walk through a refactor safely, step by step

Docs & Tracking

• 📝 /update-docs — Keep documentation in sync with your code
• 📚 /gather-tech-docs — Pull together all the technical documentation in a project
• 📓 /dev-diary — Track decisions and context as you build
• 🚀 /post-init-onboarding — Get oriented in a brand new codebase quickly

Testing & Quality

• 🧪 /tdd — Write your tests first, then build to make them pass
• 🎨 /visual-test — Check for visual regressions using Playwright
• 🔒 /security-audit — Scan for common security vulnerabilities
• ⚡ /performance-check — Find the slow parts of your code
• 💸 /tech-debt-hunt — Surface the parts of the codebase that most need attention
• 🔬 /ultrareview — Run a thorough multi-angle review of everything on your branch

---

🪝 Background Automations (Hooks)

Hooks are little scripts that run automatically when Claude does something — like a formatter that fires every time Claude edits a file, or a notification that pops up when a long task finishes.

Two hooks come included:

Hook	When it runs	What it does
Auto-format	Every time Claude edits a file	Runs Prettier, Ruff, gofmt, or rustfmt depending on the file type
Task done	When Claude finishes working	Sends you a desktop notification so you can step away without babysitting

These activate automatically when setup.sh installs ~/.claude/settings.json. If you already have that file, check out docs/hooks-guide.md to add them manually.

You can also write your own hooks — block certain commands, log everything Claude touches, post to Slack when something finishes. The guide covers all of it.

---

🧠 Scoped Rules — Claude Knows Your Style

Scoped rules are instructions for Claude that only activate when it's working on a specific type of file. So your TypeScript rules load when Claude opens a .ts file, your Python rules load for .py files, and so on — without cluttering up every conversation.

Three rules come built in:

Rule	Activates for	What it covers
TypeScript	.ts, .tsx, .js, .jsx	Type safety, async best practices, things to avoid
Python	.py	Type hints, modern Python patterns, formatting
Testing	Test files	How to write reliable, maintainable tests

Where they live:

• ~/.claude/rules/ — installed by setup.sh, apply across all your projects
• .claude/rules/ in a repo — apply only to that project

To add your own rule, create a file with this at the top:

---
globs: "**/*.ext"
---
Your rules go here.

---

🤖 MCP Plugins — Connect Claude to More Tools

MCP plugins extend what Claude can do. The great thing is they only "wake up" when you actually use them — so having them installed doesn't slow anything down.

Plugin	What it gives Claude
Playwright	Control a real browser — great for testing UI
Context7	Up-to-date docs for any library or framework
Filesystem	Structured access to your files (optional, manual setup)
GitHub	Read and write issues, PRs, and code on GitHub (needs a token)
Sequential Thinking	Step-by-step structured reasoning for complex problems

setup.sh handles Playwright and Context7 automatically. The others need a quick manual setup — copy config/mcp.json to ~/.claude/mcp.json and fill in your paths and tokens (there are notes in the file explaining each one).

---

🛠️ Shell Helpers — For Claude's Use (and Yours)

These are short commands that Claude uses behind the scenes to work faster in your project. You can run them yourself too!

• chp — Instant project overview. Run this first in any new project.
• chs find-code "something" — Search your codebase fast
• ch — Access any helper: ch [category] [command]

Category	Alias	What it does
project	p	Full project overview (chp)
search	s	Fast code and file search
git	g	Streamlined git workflows
docker	d	Container management
typescript	ts	Node.js/TypeScript tools
python	py	Python environment and testing
go	go	Go modules and testing
context	ctx	Generate focused context for a specific task
multi	m	Read or process multiple files at once
api	—	Test HTTP endpoints
interactive	i	Pick files and branches interactively
nlp	text	Code analysis: complexity, security, duplication

Run ch [category] help to see everything in any category.

---

🔍 Code Analysis — Spot Problems Before They Ship

The ch nlp tools let Claude (or you) analyze code without loading everything into the conversation. Run them before a PR to catch issues early.

ch nlp overview app.py      # Full picture: complexity, security, quality, docs
ch nlp security auth.py     # Check for accidental secrets or injection risks
ch nlp complexity utils.js  # Find functions that are getting too tangled
ch nlp tokens bigfile.ts    # See how large a file is before loading it

No extra tools needed — everything runs with Python's built-in libraries.

---

🔧 Required & Optional Tools

Required (most systems already have these):

• ripgrep — fast searching
• jq — JSON processing

Optional (setup script will offer to install):

• fzf — interactive file and branch picker
• bat — syntax-highlighted file viewing
• gum — nice-looking CLI prompts
• delta — better git diffs
• httpie — friendlier HTTP client

---

✏️ Make It Yours

Add a new slash command:

1. Create commands/my-command.md with instructions for Claude
2. Run ./setup.sh
3. Use it as /my-command in Claude Code

Add a scoped rule:

1. Create .claude/rules/my-rule.md
2. Add ---\nglobs: "**/*.ext"\n--- at the top
3. Write your rules — Claude picks them up automatically

Add a helper script:

1. Create scripts/my-helper.sh
2. Run ./setup.sh
3. Access it via ch myhelper

---

🤝 Contributing

Found a workflow that saves you time? A hook that's been useful? Add it and share it!

1. Fork the repo
2. Add your scripts or commands
3. Open a pull request

---

📄 License

MIT License — see LICENSE for details