Claude Code Command Collection

A collection of powerful, specialized commands for Claude Code - Anthropic's official CLI for Claude.

Each command is a self-contained package that can be installed independently or all together using the installer.

Quick Start

Option 1: Clone Everything (Recommended)

# Clone with all commands
git clone --recursive https://github.com/tx2z/claude-code-commands.git

# Install commands to your project using the interactive installer
cd claude-code-commands
./install.sh --interactive

# Or install all commands at once
./install.sh --all -d /path/to/your-project

Option 2: Install Specific Commands

# Clone the collection
git clone --recursive https://github.com/tx2z/claude-code-commands.git
cd claude-code-commands

# Install only what you need
./install.sh security-scan code-review -d /path/to/your-project

Option 3: Clone Individual Commands

Each command has its own repository for standalone installation:

cd your-project
mkdir -p .claude/commands
git clone https://github.com/tx2z/claude-code-security-scan .claude/commands/security-scan

Installer Usage

./install.sh [OPTIONS] [COMMAND...]

Options:
  -h, --help          Show help message
  -l, --list          List available commands
  -a, --all           Install all commands
  -i, --interactive   Interactive mode (select commands)
  -d, --directory     Target directory (default: current directory)

Examples:
  ./install.sh --interactive                    # Interactive selection
  ./install.sh --all                            # Install all to current dir
  ./install.sh --all -d ~/projects/myapp        # Install all to specific dir
  ./install.sh security-scan code-review        # Install specific commands

Personalizing Commands (Optional)

After installation, the script creates a .claude/PERSONALIZE.md file with prompts you can run in Claude Code to optimize each command for your specific tech stack.

Why personalize?

• Removes irrelevant patterns (e.g., Django checks from a Node.js project)
• Adds project-specific conventions and patterns
• Improves accuracy and reduces scan time

How to personalize:

1. Open your project in Claude Code
2. Copy a prompt from .claude/PERSONALIZE.md
3. Paste it into Claude Code and run it
4. Review the proposed changes before applying

This is optional but recommended for best results.

Available Commands

Command	Description	Repository
/security-scan	OWASP security vulnerability scanner	claude-code-security-scan
/code-review	Multi-perspective code reviews (Peer, Architect, Security, Product, CTO)	claude-code-code-review
/refactor-scan	Identify refactoring opportunities and technical debt	claude-code-refactor-scan
/doc-gen	Generate documentation (README, API docs, architecture)	claude-code-doc-gen
/api-validator	Validate APIs against REST/GraphQL best practices	claude-code-api-validator
/test-coverage-audit	Audit test quality beyond line coverage	claude-code-test-coverage-audit
/performance-scan	Detect performance issues and optimization opportunities	claude-code-performance-scan
/release-notes	Generate release notes in multiple formats	claude-code-release-notes
/init-kb /update-kb	Generate and maintain project knowledge base	claude-code-kb-generator

Command Overview

/security-scan

Comprehensive OWASP security scanning with 11 specialized agents covering:

• OWASP Top 10 (Web)
• OWASP API Security Top 10
• Secret detection
• License compliance
• CVE vulnerability checking
• Docker security

Modes: full, quick, api-only, web-only, secrets-only, category:A01-A10, category:API1-API10

---

/code-review

Multi-perspective code reviews using 5 specialized agents:

• PEER01 - Peer Developer (readability, DRY, SOLID)
• ARCH01 - Software Architect (patterns, coupling, scalability)
• SEC01 - Security Reviewer (OWASP, vulnerabilities)
• PROD01 - Product Perspective (UX, a11y, i18n)
• CTO01 - CTO/Strategic (maintainability, compliance, cost)

Modes: full, quick, peer, arch, security, product, cto, pr

---

/refactor-scan

Identify refactoring opportunities with 6 agents:

• REF01 - Code Complexity (cyclomatic, nesting, method length)
• REF02 - Duplication (copy-paste, DRY violations)
• REF03 - Design Patterns (factory, strategy, observer opportunities)
• REF04 - Code Smells (Martin Fowler's catalog)
• REF05 - Modernization (outdated patterns, upgrades)
• REF06 - Architecture Debt (circular deps, god objects)

Modes: full, quick, complexity, duplication, patterns, modernize, debt, extract

---

/doc-gen

Generate comprehensive documentation with 6 agents:

• DOC01 - README Generator
• DOC02 - API Documentation (REST/GraphQL)
• DOC03 - Architecture Documentation (diagrams, ADRs)
• DOC04 - Setup Guide
• DOC05 - Inline Comments (JSDoc, docstrings)
• DOC06 - Component Documentation (React/Vue)

Modes: full, readme, api, arch, setup, contrib, changelog, inline, storybook

---

/api-validator

Validate APIs against best practices with 7 agents:

• API01 - REST Best Practices
• API02 - GraphQL Best Practices
• API03 - OpenAPI/Swagger Spec
• API04 - API Security
• API05 - API Performance
• API06 - API Consistency
• API07 - API Versioning

Modes: full, rest, graphql, openapi, security, performance, versioning, consistency

---

/test-coverage-audit

Audit test quality with 6 agents:

• TEST01 - Coverage Gaps
• TEST02 - Test Quality (AAA pattern, isolation)
• TEST03 - Mutation Testing Analysis
• TEST04 - Flaky Tests Detection
• TEST05 - Test Performance
• TEST06 - Test Organization

Modes: full, gaps, quality, unit, integration, e2e, mutation, flaky, performance

---

/performance-scan

Detect performance issues with 7 agents:

• PERF01 - Algorithm Complexity
• PERF02 - Memory Issues
• PERF03 - Database Performance
• PERF04 - Frontend Performance
• PERF05 - Network Optimization
• PERF06 - Async Operations
• PERF07 - Build/Bundle Analysis

Modes: full, quick, frontend, backend, database, memory, bundle, runtime, network

---

/release-notes

Generate release notes in multiple formats with 5 agents:

• REL01 - Technical Notes (GitHub releases, changelogs)
• REL02 - Marketing Notes (App Store, user-friendly)
• REL03 - Internal Notes (team communication)
• REL04 - Security Notes (CVE disclosures)
• REL05 - Changelog Maintenance

Modes: technical, marketing, internal, security, full

Arguments: --from=<tag>, --to=<tag>, --version=<semver>, --date=<date>

---

/init-kb & /update-kb

Generate and maintain a comprehensive knowledge base for your project:

• Automated CLAUDE.md generation
• Project structure analysis
• Dependency documentation
• Architecture insights

Commands: /init-kb (initial generation), /update-kb (incremental updates)

---

Tech Stack Support

All commands support multiple programming languages and frameworks:

• TypeScript/JavaScript - React, Node.js, Next.js, Express, Vue, Angular
• Python - Django, FastAPI, Flask
• PHP - Laravel, Symfony
• .NET/C# - ASP.NET Core, Entity Framework
• Go - Gin, Echo, Chi
• Java - Spring Boot
• Rust

Updating Commands

To update all commands to their latest versions:

cd claude-code-commands
git submodule update --remote --merge

To update a specific command:

cd claude-code-commands/claude-code-security-scan
git pull origin main

Contributing

Each command is maintained in its own repository. Please submit issues and pull requests to the specific command repository.

License

All commands are released under the MIT License.

Author

Created by tx2z