AI Rules Template

A comprehensive rules template for AI coding assistants, supporting Cursor AI and Claude Code with Context Engineering.

Overview

This template provides a structured framework for AI-assisted development based on software engineering best practices and documentation-driven development. It helps AI coding assistants understand project context, follow conventions, and maintain code quality.

Features

• Cursor AI Support: .cursor/rules/ directory with .mdc files for on-demand loading
• Claude Code Support: Context Engineering workflow with PRP (Product Requirements Prompts)
• Memory Bank: Persistent documentation system in docs/ and tasks/
• Agile Workflow: Built on software engineering principles and best practices
• Auto-Documentation: Automatically updates documentation after changes

Quick Start

Cursor AI

1. Copy .cursor/rules/ to your project root
2. Start coding with Cursor AI

Claude Code

1. Copy .claude/ directory and CLAUDE.md to your project root
2. Install Claude Code from Anthropic
3. Use PRP workflow:
   • Write requirements in INITIAL.md
   • Run /generate-prp INITIAL.md
   • Run /execute-prp PRPs/your-feature-name.md

Directory Structure

project/
├── .cursor/rules/      # Cursor AI rules
├── .claude/            # Claude Code config
├── CLAUDE.md           # Claude Code global rules
├── AGENTS.md           # MiMoCode rules
├── MEMORY.md           # Memory documentation system
├── docs/               # Documentation
│   ├── architecture.md
│   ├── technical.md
│   └── product_requirement_docs.md
├── tasks/              # Task management (general project tasks)
│   ├── active_context.md
│   ├── tasks_plan.md
│   └── changelog.md
├── PRPs/               # Product Requirements Prompts (Claude Code workflow)
│   ├── templates/
│   │   └── prp_base.md
│   └── EXAMPLE_multi_agent_prp.md
├── examples/           # Code examples
└── src/                # Source code

tasks/ vs PRPs/

• tasks/: General project task management for all AI assistants

  • active_context.md: Current development context and focus
  • tasks_plan.md: Task backlog and project progress
  • changelog.md: Change history and documentation

• PRPs/: Claude Code-specific Product Requirements Prompts

  • templates/prp_base.md: PRP template with validation loops
  • EXAMPLE_multi_agent_prp.md: Complete PRP example
  • Used with /generate-prp and /execute-prp commands

Rule Files

Cursor AI

.cursor/rules/
├── rules.mdc           # Core rules
├── plan.mdc            # Planning workflow
├── implement.mdc       # Implementation workflow
├── debug.mdc           # Debugging workflow
├── memory.mdc          # Documentation system
└── directory-structure.mdc

Memory System

MEMORY.md provides a comprehensive documentation system for persistent project memory:

• Core Files (required): Product requirements, architecture, technical specs, task plans, active context, error documentation, lessons learned
• Context Files (optional): Literature surveys, RFCs
• Workflow: PLAN/ACT modes for reading and writing memory files
• Auto-Updates: Documentation automatically updated after changes

Claude Code

.claude/
├── commands/
│   ├── generate-prp.md
│   └── execute-prp.md
└── settings.local.json

Core Principles

1. Context Engineering: Provide comprehensive context for AI assistants
2. Validation Loops: Self-correcting workflows with executable tests
3. Documentation-Driven: Persistent memory through structured documentation
4. Modular Design: Separation of concerns and incremental development

Contributing

See CONTRIBUTING.md for guidelines.

Acknowledgments

This project is inspired by and references:

• rules_template by Bhartendu Kumar
• context-engineering-intro by Cole McMahan

License

MIT