CoachFinder Claude Code Toolkit

A comprehensive Claude Code toolkit for building high-quality, secure, and well-tested SaaS applications. Designed for teams ranging from early-career developers to non-technical founders.

What This Toolkit Provides

• AGENTS.md: Always-loaded development guidance with decision routing
• Skills: On-demand tools for security review, PR review, testing, and more
• Best Practices: Reference documentation for coding standards
• Persona Guides: Tailored guides for developers and founders

Quick Start

Installation

1. Clone this repository into your project:

   git clone https://github.com/CoachCoe/CoachFinderCC.git .claude-toolkit

2. Run the setup script:

   ./claude-toolkit/scripts/setup.sh /path/to/your/project

   Or manually copy the files:

   cp .claude-toolkit/AGENTS.md /path/to/your/project/
   cp .claude-toolkit/CLAUDE.md /path/to/your/project/
   cp -r .claude-toolkit/.claude /path/to/your/project/
   cp -r .claude-toolkit/docs /path/to/your/project/

3. Customize CLAUDE.md for your project (see checklist below)

Customization Checklist

After installation, open CLAUDE.md and update these sections:

• Project Overview - Describe what your app does in 2-3 sentences
• Repository Structure - Run tree -L 2 -d (or find . -type d -maxdepth 2) and paste your actual folder structure
• Development Commands - Check your package.json scripts and update the commands
• Environment Variables - List the variables from your .env.example (names only, never values)
• Key Patterns - Document how auth works, your data models, naming conventions
• Known Technical Debt - Add any shortcuts or issues you're aware of
• Questions to Investigate - Note anything about the codebase you don't fully understand

Tip: You don't have to do this all at once. Start with Repository Structure and Development Commands, then fill in the rest as you learn the codebase.

Verification

After installation, verify the toolkit is working:

# In Claude Code
/pr-review

You should see the PR Review skill activate and analyze your repository.

---

Available Skills

Skill	Command	Description
PR Review	/pr-review	Comprehensive PR review for quality, security, and testing
Security Review	/security-review	Analyze code for security vulnerabilities
Audit Prep	/audit-prep	Pre-launch readiness checklist
Code Quality	/code-quality	Check for bloat, over-engineering, best practices
Test Quality	/test-quality	Verify tests are meaningful + mutation testing
Investigate Component	/investigate-component	Document unknown code components
Skill Creator	/skill-creator	Create new skills for your project

---

For Developers

See docs/FOR-DEVELOPERS.md for:

• How to use skills effectively
• Working with Claude Code
• Quality standards
• Common patterns
• Growing your skills

Quick Reference

# Before merging any PR
/pr-review

# For security-sensitive changes
/security-review

# Before launch
/audit-prep

# Check test quality
/test-quality

# Understand unknown code
/investigate-component "feature name"

---

For Founders

See docs/FOR-FOUNDERS.md for:

• How to review PRs without technical expertise
• Understanding risk levels
• Launch readiness checklists
• When to involve your developer

Quick Reference

When You Want To...	Say This
Review a PR	/pr-review
Check launch readiness	/audit-prep
Understand something	/investigate-component X
Check security	/security-review

---

Architecture

your-project/
├── AGENTS.md                 # Always-loaded guidance
├── CLAUDE.md                 # Project-specific context
├── .claude/
│   ├── settings.json         # Permissions configuration
│   └── skills/
│       ├── security-review/  # Security analysis skill
│       ├── audit-prep/       # Launch readiness skill
│       ├── code-quality/     # Code review skill
│       ├── test-quality/     # Test verification skill
│       ├── pr-review/        # Combined PR review skill
│       ├── investigate-component/  # Code investigation skill
│       └── skill-creator/    # Meta-skill for creating skills
└── docs/
    ├── FOR-DEVELOPERS.md     # Developer guide
    ├── FOR-FOUNDERS.md       # Founder guide
    └── BEST-PRACTICES.md     # Coding standards reference

How It Works

1. AGENTS.md is always loaded into Claude Code's context, providing baseline guidance and decision routing.

2. Skills are invoked on-demand when you use a command like /security-review. They provide structured workflows for specific tasks.

3. CLAUDE.md contains your project-specific context that Claude uses to understand your codebase.

4. settings.json configures what operations Claude Code can perform automatically vs. what requires your approval.

---

Customization

Adding Project-Specific Context

Edit CLAUDE.md to include:

• Your specific tech stack details
• Project conventions
• Known technical debt
• Questions to investigate

Creating New Skills

Use /skill-creator to create skills specific to your project:

/skill-creator

Follow the template to define:

• When the skill should activate
• What steps it should follow
• What output it should produce

Adjusting Permissions

Edit .claude/settings.json to customize what Claude Code can do automatically:

{
  "permissions": {
    "allow": ["Read", "Glob", "Bash(pnpm:*)"],
    "ask": ["Bash(git push:*)"],
    "deny": ["Bash(rm -rf /*)"]
  }
}

---

Tech Stack Support

This toolkit is optimized for:

• Frontend: Next.js (App Router), React, Vite
• Language: TypeScript (strict mode)
• Backend: Supabase (Auth, Database, Storage, Realtime)
• Payments: Stripe
• Styling: Tailwind CSS

The principles and patterns apply broadly to any web application, but the specific examples and security checks are tailored for this stack.

---

Contributing

To contribute to this toolkit:

1. Fork the repository
2. Create a new skill or improve existing ones
3. Test your changes
4. Submit a pull request

See docs/FOR-DEVELOPERS.md for coding standards.

---

License

MIT License - See LICENSE file for details.

---

Support

• Issues: Report bugs or request features via GitHub Issues
• Questions: Ask Claude Code directly - it knows this toolkit

---

Acknowledgments

This toolkit was built based on research and best practices from:

• Agent Skills for Context Engineering
• Vercel AGENTS.md Research
• Stryker Mutation Testing
• Claude Code best practices and patterns