specOS is a comprehensive framework for Spec Driven AI-assisted software development, designed to work seamlessly with Claude Code and other AI development tools.
specOS provides a systematic approach to:
- Product Planning: Structured product ideation and roadmap development
- Feature Specification: Detailed, testable feature specifications
- Task Execution: Systematic implementation with TDD approach
- Code Quality: Automated code review and bug fixing
- Project Management: Git workflow automation and documentation
specOS/
├── claude-code/ # Claude Code integration
├── agents/ # Specialized AI agents
├── user/ # User configuration
├── commands/ # Claude Code commands
├── instructions/ # Workflow instructions
├── standards/ # Development standards
└── setup.sh # Installation script
- bug-fixer: Systematic debugging and error resolution
- code-reviewer: Comprehensive code quality and security review
- data-scientist: SQL analysis and data insights
- context-fetcher: Smart information retrieval
- file-creator: Template-based file generation
- git-workflow: Automated git operations
Run this command in your project directory to install specOS:
curl -sSL https://raw.githubusercontent.com/rainday/specOS/main/setup.sh | bashWhat this does:
- Downloads and executes the specOS setup script
- Creates necessary directories in your home folder (
~/.specOS/,~/.claude/) - Downloads all development standards, instructions, and Claude Code configurations
- Sets up specialized AI agents for code review, bug fixing, data analysis, and more
- Configures Claude Code commands for seamless integration
Alternatively, you can clone the repository and run the setup manually:
# Clone the repository
git clone https://github.com/rainday/specOS.git
cd specOS
# Run the setup script
./setup.sh-
Plan a new product:
/plan-product
-
Analyze existing codebase:
/analyze-product
-
Create a feature spec:
/create-spec
-
Execute tasks:
/execute-tasks
-
Install specOS (if not already done):
curl -sSL https://raw.githubusercontent.com/rainday/specOS/main/setup.sh | bash -
Initialize your project:
/plan-product
This will guide you through:
- Defining your product mission and vision
- Identifying target users and key features
- Setting up your tech stack
- Creating initial project structure
- Analyze your codebase:
This will:
/analyze-product
- Review your existing code structure
- Identify areas for improvement
- Suggest optimizations
- Create project documentation
-
Create a feature specification:
/create-spec
Or simply ask: "What's next?" or "Create a spec for [feature name]"
-
Implement the feature:
/execute-task
This follows TDD approach with:
- Test creation
- Implementation
- Code review
- Quality assurance
specOS includes several specialized AI agents that activate automatically:
- 🪲 Bug Fixer: Automatically detects and fixes code issues
- 🧐 Code Reviewer: Reviews code for quality, security, and best practices
- 📊 Data Scientist: Analyzes SQL queries and provides data insights
- 🔍 Context Fetcher: Retrieves relevant information from your codebase
- 📁 File Creator: Generates files from templates
- 🌳 Git Workflow: Manages git operations and commits
- Start with planning: Always use
/plan-productfor new projects - Be specific: When creating specs, provide clear requirements
- Iterate: Use the agents to review and improve your code
- Document: Let specOS handle documentation automatically
- Test first: The framework follows TDD principles
| Command | Purpose | When to Use |
|---|---|---|
/plan-product |
Initialize new project | Starting a new project |
/analyze-product |
Analyze existing codebase | Working with existing code |
/create-spec |
Create feature specification | Adding new features |
/execute-task |
Implement features | Building functionality |
What's next? |
Get next steps | General guidance |
- Define product mission and vision
- Identify target users and key features
- Create technical roadmap
- Establish development standards
- Create detailed feature specifications
- Define technical requirements
- Plan database and API changes
- Break down into testable tasks
- Follow TDD approach
- Use specialized agents for quality assurance
- Maintain consistent code standards
- Automate git workflow
- Automated code review
- Security analysis
- Performance optimization
- Comprehensive testing
- Tech Stack:
~/.specOS/standards/tech-stack.md - Code Style:
~/.specOS/standards/code-style.md - Best Practices:
~/.specOS/standards/best-practices.md
- Product Mission:
.specOS/product/mission.md - Tech Stack:
.specOS/product/tech-stack.md - Requirements:
.specOS/product/requirements.md
- plan-product.md: New product initialization
- create-spec.md: Feature specification creation
- execute-tasks.md: Task implementation workflow
- analyze-product.md: Existing codebase analysis
- tech-stack.md: Technology stack defaults
- best-practices.md: Development guidelines
- code-style/: Language-specific style guides
- Create agent file in
claude-code/agents/ - Define agent metadata and capabilities
- Update setup script to include new agent
- Edit files in
standards/directory - Run setup script with
--overwrite-standards - Test changes in a sample project
- Modify instruction files in
instructions/ - Update command files in
commands/ - Test workflow changes
- Fork the repository
- Create a feature branch
- Make your changes
- Test thoroughly
- Submit a pull request
This project is licensed under the MIT License - see the LICENSE file for details.
- Documentation: GitHub README
- Issues: GitHub Issues
- Discussions: GitHub Discussions
- v1.1.1 (2024-01-16): Added Japanese language support and improved Cursor integration
- v1.1.0 (2024-01-15): Enhanced spec creation workflow
- v1.0.0 (2024-01-10): Initial release with core functionality
Built with ?��? by the Rainday