ClaudePreference
May 21, 2026 · View on GitHub
English | 中文
A comprehensive collection of development workflow commands for Claude Code, designed to streamline software development processes from planning to deployment.
🚀 Overview
ClaudePreference provides a suite of specialized commands that integrate seamlessly with Claude Code to provide intelligent assistance for:
- Development Lifecycle Management: From requirements analysis to deployment
- Code Quality Assurance: Automated reviews, testing, and security scanning
- Project Maintenance: Cleanup, documentation, and branch management
- Architecture Planning: Strategic design decisions and code analysis
- Codex Skills: Reusable workflows for code explanation, review reports, local runtime setup, browser flow debugging, and image conversion
📋 Available Commands
💡 Complete Documentation: Visit Command Documentation Hub for detailed descriptions, usage examples, and best practices for all commands.
Core Development Workflow
1. Orchestrated Development - /m:orchestrated-dev
Multi-agent development workflow with coordinated dev-review cycles.
Usage: Deploy three specialized agents (Orchestrator, Developer, Reviewer) for collaborative development in the current directory.
Features:
-
Research-driven architecture planning with MCP tools integration
-
Template-based prompt generation for consistent agent communication
-
Dynamic context-aware prompts with strict quality enforcement
-
Build verification with 100% functional coverage requirement
-
Evidence-based decision making with external validation
-
Detailed Documentation: m:orchestrated-dev
2. Commit & Push - /m:commit-push
Automated commit and push workflow with intelligent message generation.
Parameters: [message] (optional)
Examples:
/m:commit-push "feat: add user authentication"/m:commit-push "fix: resolve memory leak"/m:commit-push(auto-generates conventional commit message)
Detailed Documentation: m-commit-push
3. Bug Fix Workflow - /m:bug-fix
Comprehensive bug analysis, reproduction, and fix workflow.
Parameters: [source] (optional) - issue number, error text, screenshot, log file, or description
Examples:
/m:bug-fix #123- Fix GitHub issue/m:bug-fix "NullPointerException in login"/m:bug-fix screenshot.png- Fix from error screenshot/m:bug-fix logs/error.log- Analyze log file
Detailed Documentation: m:bug-fix
Code Quality Management
4. Security Scan - /m:security-scan
Comprehensive security vulnerability scanning and assessment.
Parameters: [scope] (optional) - dependencies, auth, data-handling, full, or specific paths
Examples:
/m:security-scan dependencies- Scan only dependencies/m:security-scan auth- Focus on authentication security/m:security-scan src/api- Scan specific directory
Detailed Documentation: m:security-scan
5. Test Generation - /m:test-generation
Automated test case generation with coverage analysis.
Parameters: [type] [target] [coverage] (optional)
Examples:
/m:test-generation unit src/auth 90%- Generate unit tests with 90% coverage/m:test-generation integration api- Generate integration tests/m:test-generation e2e user-flow- Generate end-to-end tests
Detailed Documentation: m:test-generation Detailed Documentation:
6. Code Review - /m:review-code
Tactical code review with quality assessment and reporting.
Parameters: [target] [depth] [focus] (optional)
Examples:
/m:review-code src/auth deep security- Deep security review/m:review-code PR#123 quick- Quick PR review/m:review-code components standard performance- Performance review
Project Management & Planning
7. Task Planner - /m:task-planner
Analyze requirements and generate structured implementation plans.
Parameters: [target] (optional) - requirement documents or project scope
Features:
- Requirements analysis and task decomposition
- Dependency mapping and sequencing
- Implementation strategy design
- Visual task flow diagrams
Detailed Documentation: m:task-planner
8. TDD Planner - /m:tdd-planner
Generate comprehensive Test-Driven Development plans.
Parameters: [target] (optional) - requirement documents
Features:
- Test strategy design and framework selection
- Red-Green-Refactor cycle planning
- Test pyramid structure definition
- JSON-formatted implementation plan
Detailed Documentation: m:tdd-planner
9. Next Tasks - /m:next-task
Analyze current development status and generate prioritized action plans.
Parameters: [priority] (optional) - high, medium, low
Features:
- Documentation and tracker scanning
- Git history analysis for pending work
- Priority ranking and dependency assessment
- Comprehensive status reporting
Detailed Documentation: m:next-task
10. Next Context - /m:next-context
Compile comprehensive task lists and background information.
Parameters: [target] (optional) - project scope or conversation context
Features:
- Project background analysis
- Task extraction and classification
- Supporting material collection
- Ambiguity detection and clarification
Detailed Documentation: m:next-context
Maintenance & Documentation
11. Project Cleanup - /m:project-cleanup
Automated project hygiene and maintenance workflow.
Parameters: [scope] (optional) - code, dependencies, structure, artifacts, config, all
Features:
- Dead code elimination
- Dependency pruning and updates
- Codebase formatting and linting
- Structural reorganization
Detailed Documentation: m:project-cleanup
12. Document Update - /m:document-update
Automated documentation maintenance and updates.
Parameters: [scope] (optional) - api, readme, changelog, comments, all, or specific paths
Features:
- API documentation refresh
- README and changelog updates
- Code comment validation
- Documentation consistency checks
Detailed Documentation: m:document-update
13. Branch Cleanup - /m:branch-prune
Comprehensive branch management and cleanup workflow.
Features:
- Branch analysis and documentation
- Merged branch cleanup
- Functionality testing across versions
- Sequential merge operations
Detailed Documentation: m:branch-prune
14. Branch Cleanup Lite - /m:branch-prune-lite
Lightweight branch cleanup without testing.
Parameters: [target_branch] (optional) - specific branch to prune
Features:
- Individual branch analysis
- Merged branch cleanup
- Specific worktree and branch removal
- Sequential merge operations
Detailed Documentation: m:branch-prune-lite
Architecture & Design
15. Architecture Debate - /m:debate-architecture
Strategic architecture review and analysis workflow.
Parameters: [target] (optional) - modules, directories, or "all"
Features:
- System-wide pattern identification
- Structural design issue detection
- Strategic "-ilities" evaluation
- High-level refactoring proposals
Detailed Documentation: m:debate-architecture Detailed Documentation:
16. Code Debate - /m:debate-code
Tactical code review with detailed implementation analysis.
Parameters: [target] [depth] [focus] (optional)
Features:
- Code implementation analysis
- Implementation-level issue identification
- Coding standards compliance verification
- Line-level feedback generation
Detailed Documentation: m:debate-code
17. Review Completion - /m:review-completion
Review completion workflow and verification.
Features:
- Completion status validation
- Quality gate verification
- Deliverable confirmation
- Final review sign-off
Detailed Documentation: m:review-completion
18. Branch Merge - /m:merge-branch
Intelligent branch merging workflow.
Features:
- Smart merge strategy selection
- Conflict resolution assistance
- Post-merge validation
- Automated cleanup
Detailed Documentation: m:merge-branch
🧩 Available Codex Skills
The skills/ directory contains general-purpose Codex skills that are safe to publish and reuse across projects:
- develop-review-gate: Implement changes in the current checkout, run two self-review/refactor passes, then pause for human confirmation.
- doc-code-review-report: Review requirements, PR text, or design notes against the actual codebase and produce evidence-backed findings.
- refactor-design-report: Turn concrete bugs, product gaps, or architecture concerns into an executable implementation design.
- explain: Explain files, symbols, snippets, or configuration in repository context.
- local-project-runtime: Stabilize local build/start/test/Compose workflows after clone or checkout.
- browser-flow-fallbacks: Diagnose flaky login, OAuth, signup, and multi-step browser automation flows.
- image-converter: Convert common local image formats on macOS with installed tools.
Install them into Codex manually:
mkdir -p "${CODEX_HOME:-$HOME/.codex}/skills"
cp -R skills/* "${CODEX_HOME:-$HOME/.codex}/skills/"
🔧 Installation & Setup
Prerequisites:
- Claude Code CLI installed and configured
- OpenAI Codex CLI installed and configured if you want to use the
skills/directory - Appropriate development environment setup
- Appropriate MCP environment setup (e.g.
context7)
Automated Installation (Recommended)
Quick Install:
# Clone and install in one command
git clone https://github.com/penwyp/ClaudePreference.git
cd ClaudePreference
./install.sh
Installation Options:
# Install to default location (~/.claude/commands/m)
./install.sh
# Install commands and bundled Codex skills
./install.sh --with-skills
# Install only bundled Codex skills
./install.sh --skills-only
# Install to custom directory
./install.sh --dir /custom/path
# Preview installation without making changes
./install.sh --dry-run
# Force overwrite existing files
./install.sh --force
# Update existing installation
./install.sh --update
# Show all available options
./install.sh --help
Manual Installation
# Clone repository
git clone https://github.com/penwyp/ClaudePreference.git
cd ClaudePreference
# Copy commands manually
cp commands/* ~/.claude/commands/
Verify Installation
# Start Claude Code in any project directory
claude
# Type "/" to see available commands
# Look for commands in the "m:" directory (e.g., /m:orchestrated-dev)
Management Commands
# Update to latest version
./install.sh --update
# Remove ClaudePreference commands
./install.sh --uninstall
# Remove bundled Codex skills
./install.sh --uninstall --skills-only
# Rollback to previous version
./install.sh --rollback
📖 Usage Examples
Daily Development Workflow
# Start with task planning
/m:task-planner requirements.md
# Dev-Review Cycle
/m:orchestrated-dev requirements.md
# Generate tests for new features
/m:test-generation unit src/newfeature 85%
# Run security scan
/m:security-scan
# Commit and push changes
/m:commit-push "feat: implement user dashboard"
Bug Fix Workflow
# Analyze and fix bug
/m:bug-fix #456
# Generate regression tests
/m:test-generation unit src/bugfix
# Review and commit fix
/m:review-code src/bugfix standard
/m:commit-push "fix: resolve login timeout issue"
Pre-Release Workflow
# Comprehensive testing
/m:test-generation integration
# Update documentation
/m:document-update
# Security assessment
/m:security-scan full
# Clean up branches
/m:branch-prune
Architecture Review
# Strategic architecture analysis
/m:debate-architecture src/core
# Tactical code review
/m:debate-code src/api deep security
# Generate improvement plan
/m:task-planner architecture-improvements.md
🛠️ Best Practices
- Command Sequencing: Use commands in logical order for optimal results
- Parameter Usage: Leverage optional parameters for targeted operations
- Regular Maintenance: Run cleanup and security scans regularly
- Documentation: Keep documentation updated after significant changes
- Testing: Generate comprehensive tests before major releases
🚨 Troubleshooting
Common Issues
Command Not Found:
- Ensure Claude Code CLI is properly installed
- Verify you're in the correct project directory
- Check that the commands/ directory exists
Permission Errors:
- Ensure proper Git permissions
- Verify write access to project files
- Check Claude Code authentication
Build Failures:
- Run
/m:project-cleanupto resolve dependency issues - Ensure all prerequisites are installed
- Check for conflicting dependencies
Getting Help
For additional support:
- Check command-specific documentation in the
commands/directory - Review error messages for specific guidance
- Ensure your development environment meets all requirements
Project Structure
ClaudePreference/
├── README.md # Main project documentation
├── README_CN.md # Chinese documentation
├── LICENSE # MIT license
├── install.sh # Installation script
├── commands/ # Claude Code command definitions
│ ├── m-orchestrated-dev.md
│ ├── m-commit-push.md
│ ├── m-bug-fix.md
│ ├── m-security-scan.md
│ ├── m-test-generation.md
│ ├── m-review-code.md
│ ├── m-task-planner.md
│ ├── m-tdd-planner.md
│ ├── m-next-task.md
│ ├── m-next-context.md
│ ├── m-project-cleanup.md
│ ├── m-document-update.md
│ ├── m-branch-prune.md
│ ├── m-branch-prune-lite.md
│ ├── m-debate-architecture.md
│ ├── m-debate-code.md
│ ├── m-review-completion.md
│ └── m-merge-branch.md
├── skills/ # General-purpose Codex skills
│ ├── browser-flow-fallbacks/
│ ├── develop-review-gate/
│ ├── doc-code-review-report/
│ ├── explain/
│ ├── image-converter/
│ ├── local-project-runtime/
│ └── refactor-design-report/
└── docs/ # Comprehensive documentation
├── README.md # Documentation hub
├── en/ # English documentation
│ └── commands/ # Detailed command guides
│ └── ... # 18 command documentation files
└── zh/ # Chinese documentation
└── commands/ # Chinese command guides
└── ... # 18 command documentation files
🤝 Contributing
This project is designed to evolve with your development needs. Feel free to:
- Customize commands for your specific workflows
- Add new commands following the established patterns
- Improve existing command documentation
- Share your workflow optimizations
📄 License
This project is open source and available under standard software development practices.
Built for developers who value intelligent automation and comprehensive workflow management.