Cursor AI代码编辑器

Getting Started 🚀

1. Clone this repository or download the rules
2. Place the .mdc files in your project's .cursor/rules/ directory
3. Start Cursor and begin interacting with the AI
4. The rules will automatically take effect during your conversations

For a quickstart use these commands or follow the "Practical Workflow" below!

For new projects, simply run:

onboard project

For existing projects that need analysis:

analyze existing

For just setting up the rules without analysis:

setup rules

Project Setup Commands

• onboard project - Perform full project analysis and setup the AI-driven workflow
• setup rules - Set up rule files and directory structure without analyzing the codebase
• analyze existing - Generate an analysis report of your codebase without creating specifications

Command Reference 📜

Below is a comprehensive reference of all commands available through the Cursor rules system:

Development Workflow Rule

development-workflow-rule.mdc: Comprehensive development lifecycle system for task tracking, testing, and version control. - Creates and manages tasks with unique IDs and structured metadata - Maintains a central task index with status tracking - Integrates testing frameworks appropriate to the project type - Handles version control with conventional commit formatting - Ensures quality assurance through automated testing - Prevents completion of tasks that don't pass testing - Automatically validates and updates README to keep documentation in sync with code

Practical Workflow 🔄

Follow this concrete workflow to get the most out of these rules:

1. Initial Requirements Discussion 💬 - Open Cursor and describe your project/feature requirements to the AI - The AI will automatically create specification files in .cursor/specs/ based on your discussion - A central SPECS.md file will be generated or updated with links to all domain-specific specs

2. Automatic Task Planning and Management 📅 - Cursor AI automatically creates tasks based on your implementation requests - Each task receives a unique ID (e.g., TASK-2023-10-15-01) and is tracked in .cursor/TASKS.md - Tasks are automatically linked to their relevant specifications and include default acceptance criteria - The AI manages multiple tasks concurrently, allowing for efficient development planning

3. Review & Refine Specifications 🔍 - Review the generated specs files to ensure they accurately capture requirements - Continue the conversation with AI to refine specs as needed - Each refinement will update the relevant specification files - Cursor AI automatically updates task details when specifications change

4. AI-Driven Development Based on Specifications 🤖 - Cursor AI activates tasks when implementation begins - Simply ask: "Please implement the feature described in specs/auth/login.md" - The AI references the specs during implementation and tracks progress against the task - If you request a feature without specifications, the AI creates them first before implementation

5. Automatic Knowledge Capture During Development 🧠 - Cursor AI identifies and records important insights during development as learnings - The AI documents significant discoveries, patterns, and solutions in .cursor/learnings/ - Share reference documents with the AI by placing them in .cursor/docs/ - All knowledge is automatically indexed in .cursor/LEARNINGS.md and .cursor/DOCUMENTS.md for future reference

6. Automatic Documentation Updates 📝 - After successful tests, Cursor AI checks if README needs updating - Documentation is kept in sync with implementation - You can explicitly check README currency with readme check command - Updates are suggested based on implemented but undocumented features

7. Automatic Commits 💾 - As the AI makes changes to files, the git-commit-rule automatically stages and commits them - Commits follow conventional commit format (feat, fix, docs, etc.) based on the nature of the change - Commit messages are automatically generated with appropriate type, scope, and description - Code is only committed after all tests pass

8. AI-Managed Task Completion and Knowledge Preservation ✅ - When implementation is complete and tests pass, Cursor AI marks the task as done - The system updates task status, marks specs as completed, and extracts learnings - Learnings are preserved in .cursor/learnings/ with references to relevant files and tasks - The task's status is updated in .cursor/TASKS.md

9. Continuous Development Loop 🔄 - For new features: discuss requirements → AI generates specs → AI creates & manages tasks → AI implements → AI captures knowledge → AI updates documentation → automatic commits - For refinements: discuss changes → AI updates specs → AI updates tasks → AI implements changes → AI updates knowledge → AI updates documentation → automatic commits

This workflow ensures all development is specification-driven, task-organized, properly documented, and automatically committed with appropriate metadata. The integrated approach means nothing gets lost - requirements, implementation details, insights, and documentation are all preserved and linked together. You simply guide the process through conversation while Cursor AI handles the entire workflow.