Claude Code Starter Kit

A production-ready boilerplate for supercharging your development workflow with Claude Code. Includes battle-tested agents, slash commands, skills, and hooks that work with any tech stack.

Quick Start

Option 1: Zero-Install (Recommended)

Instantly initialize any new or existing project with this starter kit by running a single command inside your project directory. This works flawlessly across macOS, Linux, and Windows, and completely avoids Git SSH configuration issues:

npx @cortexguardai/claude-starter-kit

Option 2: Clone and customize

git clone https://github.com/CortexGuardAI/claude-starter-kit.git my-project
cd my-project

Option 3: Copy into existing project

# Create the .claude directory if it doesn't exist
mkdir -p /path/to/your/project/.claude

# Copy the folders into your project's .claude folder
cp -r agents/ commands/ skills/ hooks/ mcp-configs/ /path/to/your/project/.claude/

# Copy CLAUDE.md and local .claude.json to your project root
cp CLAUDE.md /path/to/your/project/
cp mcp-configs/mcp-servers.json /path/to/your/project/.claude.json

Then edit CLAUDE.md to match your project's specifics.

What's Included

Agents

Specialized sub-agents that Claude delegates to for specific tasks.

| Agent | File | Description | |-------|------|-------------| | Planner | agents/planner.md | Creates detailed implementation plans with phased steps, risk assessment, and testing strategy | | Architect | agents/architect.md | System design, trade-off analysis, ADR templates, and scalability planning | | Loop Operator | agents/loop-operator.md | Runs autonomous fix/refactor loops with safety controls, checkpoints, and escalation | | TDD Guide | agents/tdd-guide.md | Enforces Red-Green-Refactor cycle with 80%+ coverage across unit, integration, and E2E tests | | Security Reviewer | agents/security-reviewer.md | OWASP Top 10 detection, secrets scanning, input validation, and dependency auditing |

Commands

Slash commands you invoke directly (e.g., /plan, /tdd).

| Command | File | Description | |---------|------|-------------| | /plan | commands/plan.md | Analyze requirements, create step-by-step plan, wait for approval before coding | | /tdd | commands/tdd.md | Scaffold interfaces, write tests first, implement, refactor, verify coverage | | /build-fix | commands/build-fix.md | Auto-detect build system, parse errors, fix one at a time with guardrails | | /code-review | commands/code-review.md | Review uncommitted changes for security (CRITICAL), quality (HIGH), practices (MEDIUM) | | /checkpoint | commands/checkpoint.md | Create/verify git-based workflow checkpoints with state comparison | | /docs | commands/docs.md | Look up current documentation for any library via Context7 MCP | | /setup-github | commands/setup-github.md | Configures the GitHub MCP server with your Personal Access Token in .claude.json |

Skills

Deep workflow knowledge that Claude activates automatically when relevant.

| Skill | Directory | Description | |-------|-----------|-------------| | TDD Workflow | skills/tdd-workflow/ | Testing patterns, mocking strategies, coverage verification, and common mistakes | | Security Review | skills/security-review/ | Secrets management, input validation, SQL injection prevention, pre-deployment checklist | | Codebase Onboarding | skills/codebase-onboarding/ | Systematic discovery process for mapping and understanding new codebases | | Semantic Commits | skills/semantic-commits/ | Version control strategy enforcing atomic, semantic conventional commits |

MCP Configurations

Pre-configured Model Context Protocol servers to give Claude Code live access to external data. See mcp-configs/README.md for setup instructions.

| Configuration | Description | |---------------|-------------| | Context7 | Live framework documentation lookups using @upstash/context7-mcp | | Playwright | Browser automation and E2E testing using @playwright/mcp | | GitHub | Read/write PRs, issues, and repos using @modelcontextprotocol/server-github |

Hooks

Event-driven automations that run before/after Claude's tool executions.

| Hook | Event | Behavior | |------|-------|----------| | No-verify blocker | PreToolUse (Bash) | Blocks git --no-verify to protect pre-commit hooks | | File size limit | PreToolUse (Write) | Blocks creation of files > 800 lines | | TODO/FIXME warning | PreToolUse (Edit) | Warns when adding TODO/FIXME/HACK comments | | console.log warning | PostToolUse (Edit) | Warns about console.log in edited code | | Test file reminder | PostToolUse (Write) | Warns when creating source files without tests | | Console.log audit | Stop | Checks all modified files for console.log |

Project Structure

claude-starter-kit/
├── CLAUDE.md              # Project-level instructions for Claude
├── README.md              # This file
├── agents/                # Specialized sub-agents
│   ├── planner.md
│   ├── architect.md
│   ├── loop-operator.md
│   ├── tdd-guide.md
│   └── security-reviewer.md
├── commands/              # Slash commands
│   ├── plan.md
│   ├── tdd.md
│   ├── build-fix.md
│   ├── code-review.md
│   ├── checkpoint.md
│   └── docs.md
├── skills/                # Deep workflow knowledge
│   ├── tdd-workflow/
│   │   └── SKILL.md
│   ├── security-review/
│   │   └── SKILL.md
│   └── codebase-onboarding/
│       └── SKILL.md
└── hooks/                 # Event-driven automations
    ├── hooks.json
    └── README.md

File Format Reference

Agents

Markdown with YAML frontmatter:

---
name: agent-name
description: What this agent does
tools: ["Read", "Grep", "Glob", "Bash", "Edit", "Write"]
model: opus | sonnet
---

Commands

Markdown with YAML frontmatter:

---
description: What this command does when invoked
---

Skills

Markdown with YAML frontmatter in a SKILL.md file inside a named directory:

---
name: skill-name
description: When and why to use this skill
---

Hooks

JSON configuration in hooks.json:

{
  "hooks": {
    "PreToolUse": [...],
    "PostToolUse": [...],
    "Stop": [...]
  }
}

Customization Guide

Adding a New Agent

1. Create agents/your-agent.md
2. Add YAML frontmatter (name, description, tools, model)
3. Define the agent's role, workflow, and output format
4. Reference it from your commands

Adding a New Command

1. Create commands/your-command.md
2. Add YAML frontmatter (description)
3. Document usage, workflow, and examples
4. Reference related agents and skills

Adding a New Skill

1. Create skills/your-skill/SKILL.md
2. Add YAML frontmatter (name, description)
3. Include "When to Activate" section
4. Provide detailed patterns, checklists, and examples

Adding a New Hook

1. Edit hooks/hooks.json
2. Add entry to appropriate event (PreToolUse, PostToolUse, Stop)
3. Set the matcher (tool name pattern)
4. Write the hook command (receives JSON on stdin, outputs JSON on stdout)
5. Use exit code 2 to block (PreToolUse only), stderr to warn

Recommended Workflow

1. /plan          --> Create implementation plan
2. Review         --> Approve or modify the plan
3. /tdd           --> Implement with tests first
4. /checkpoint    --> Save progress
5. /build-fix     --> Fix any build errors
6. /code-review   --> Review before committing
7. /checkpoint    --> Verify final state

Contributing

1. Follow the file format conventions above
2. Keep files under 800 lines
3. Use lowercase with hyphens for filenames (e.g., my-agent.md)
4. Include examples in all agents, commands, and skills
5. Test hooks locally before adding to hooks.json

License

MIT