Please note: The new project is CursorRiper. This repo will not be maintained.
A robust, advanced rule set framework integrates three powerful AI-assisted coding strategies within the Cursor IDE. This comprehensive system is designed to temper Claude 3.7's over-enthusiasm, ensuring it stays on track and avoids veering off course.
- START PHASE The START phase is a preprocessing phase that runs once at the beginning of a new project or major component. It focuses on project initialization, scaffolding, and setting up the Memory Bank with baseline information. start-phase-readme.md
flowchart TD
Start[BEGIN START PHASE] --> Req[Step 1: Requirements Gathering]
Req --> Decision{Recommendations?}
Decision -->|Yes| AutoRecs[AI Recommendations]
Decision -->|No| Manual[Manual Process]
AutoRecs --> Scaffold[Step 4: Project Scaffolding]
Manual --> Tech[Step 2: Technology Selection]
Tech --> Arch[Step 3: Architecture Definition]
Arch --> Scaffold
Scaffold --> Setup[Step 5: Environment Setup]
Setup --> Memory[Step 6: Memory Bank Initialization]
Memory --> End[TRANSITION TO RIPER]
- RIPER-5 Mode System: A structured workflow with five distinct operational modes (Research, Innovate, Plan, Execute, Review) to ensure precise, controlled development. Ripper Framework: by robotlovehuman, github Original post on the Cursor Forum
flowchart LR
R[Research] -->|Cyclical| I[Innovate]
I --> P[Plan]
P --> E[Execute]
E --> Rev[Review]
Rev -.-> R
-
Memory Bank: A persistent documentation system that maintains project context between sessions through hierarchical memory files. Memory bank sources: file-cursor-memory-bank-rules-md .cursorrules
-
Context Management: A dynamic system for tracking relevance-scored information, personal preferences, and implementation plans.
- Strict Operational Protocol: Clear mode declarations prevent unintended modifications
- Persistent Context: Never lose important project information between sessions
- Structured Documentation: Standardized templates for all memory files
- AI-Powered Recommendations: Option to receive intelligent technology and architecture suggestions
- Sequential Process Enforcement: Step-by-step project initialization with explicit user confirmation
- Implementation Planning: Thorough planning with clarifying questions and checklists
- Automatic Updates: Memory files stay current through continuous documentation
- Project Intelligence: Learning from patterns and preferences over time
- Copy this repository into your Cursor IDE project
- Add the files to the Cursor rules within your project
- Initialize the Start phase using the provided templates (in the chat window type
/start) - Complete Step 1: Requirements Gathering
- Choose between AI recommendations or manual technology and architecture selection
- Follow the guided process to create project structure and memory bank
- Begin with RESEARCH mode to understand your codebase
- Progress through INNOVATE and PLAN modes before implementation
- Track progress and maintain documentation as you work
The framework provides a seamless integration between the START phase and the RIPER workflow:
-
START Phase: One-time preprocessing focused on project initialization
- Requirements gathering
- Technology selection (manual or AI-recommended)
- Architecture definition (manual or AI-recommended)
- Project scaffolding with explicit file creation
- Environment setup
- Memory bank initialization
-
RIPER Workflow: Cyclic process for ongoing development
- Research: Understanding the problem and codebase
- Innovate: Generating solutions and approaches
- Plan: Detailed implementation planning
- Execute: Implementation with focus and discipline
- Review: Evaluating results and planning next steps
See the full cursor_enhanced_rule_set.md file for complete documentation using this framework. You can rename the .cursorrules file and include it in your project or add the individual files to the project's ruleset.
This framework prevents coding disasters while maintaining perfect continuity across sessions.
Works without MCP Servers
Works better with MCP Servers