โ† Back to Home

๐Ÿ“– How to Use AI Agent Configuration Files

Complete guide to set up and use AI agent rules in your projects

๐Ÿ“ฅ Step 1: Download the Language Pack

  1. Go to ai.grosgaming.org
  2. Choose your programming language
  3. Click "๐Ÿ“ฅ Download" to get the .zip archive
  4. Save the file to your computer

Example: TYPESCRIPT.zip, PYTHON.zip, RUST.zip, etc.


๐Ÿ“‚ Step 2: Extract the Archive

Windows

  1. Right-click on the .zip file
  2. Select "Extract All..."
  3. Choose your project directory
  4. Click "Extract"

macOS

  1. Double-click the .zip file (auto-extracts)
  2. Move extracted folder to your project directory

Linux

unzip TYPESCRIPT.zip -d /path/to/your/project/

๐Ÿ“‹ Step 3: Files You'll Find

After extraction, you'll have 7 files:

your-project/ โ”œโ”€โ”€ .cursorrules # Quick reference for Cursor IDE โ”œโ”€โ”€ .gitignore # Language-specific ignore patterns โ”œโ”€โ”€ AGENTS_RULES.md # โญ MAIN FILE - Complete rules โ”œโ”€โ”€ AGENTS.md # Quick reference guide โ”œโ”€โ”€ README.md # Project template โ”œโ”€โ”€ CODE_QUALITY_PRINCIPLES.md # Code quality guidelines โ””โ”€โ”€ CONTRIBUTING.md # Contribution workflow

๐ŸŒŸ Most Important File

AGENTS_RULES.md is the complete documentation that AI agents must read first.


๐Ÿค– Step 4: Tell Your AI Agent to Read the Files

Option A: Using Cursor IDE (Recommended)

Cursor automatically reads .cursorrules on project load. Then say:

Please read AGENTS_RULES.md completely before we start coding.

Option B: Using GitHub Copilot Chat

@workspace Please read the following files in order:
1. AGENTS_RULES.md (complete rules)
2. AGENTS.md (quick reference)
3. CONTRIBUTING.md (workflow)

Read them completely before we start coding.

Option C: Using ChatGPT / Claude

  1. Attach the files to your conversation:
    • AGENTS_RULES.md
    • AGENTS.md
    • CONTRIBUTING.md
  2. Say:
I've attached AI agent configuration files for [LANGUAGE].
Please read AGENTS_RULES.md completely before we start coding.
This file contains:
- TDD/DDD workflow
- Language-specific best practices
- Git rules
- Testing requirements (90% coverage)
- Code quality principles

Confirm you've read and understood the rules before we proceed.

Option D: Using Any AI Assistant

Before we start coding, please read these files in my project:

1. AGENTS_RULES.md - Complete AI agent rules (READ THIS FIRST)
2. AGENTS.md - Quick reference
3. CODE_QUALITY_PRINCIPLES.md - Quality guidelines
4. CONTRIBUTING.md - Contribution workflow

Key points to remember:
- Follow TDD (Test-Driven Development)
- Write tests first (Red-Green-Refactor)
- Maintain 90% code coverage minimum
- NEVER use git stash, git push, or git push --force
- ALWAYS ask for approval before committing
- Use conventional commit format: <type>(agt): <description>

Confirm you understand before we start.

โœ… Step 5: Verify AI Agent Understanding

After the AI confirms, ask:

What are the key rules you must follow?

Expected response should include:


๐Ÿš€ Step 6: Start Coding

Once the AI confirms understanding, you can start:

Let's implement [FEATURE]. 
Remember to follow the TDD workflow from AGENTS_RULES.md.

The AI will:

  1. ๐Ÿ“‹ Plan - Discuss the approach with you
  2. ๐Ÿงช Test - Write failing tests first
  3. ๐Ÿ’ป Develop - Implement the feature
  4. โœ… Test - Run tests (90% coverage)
  5. ๐Ÿ“ Commit - Ask your approval before committing

๐Ÿ“š Quick Reference During Development

When Starting a New Feature

Let's implement [FEATURE] following TDD.
Please create a todo-list first.

When AI Forgets the Rules

Please re-read AGENTS_RULES.md, especially the [SECTION] part.

Before Committing

Review the Git Workflow section in AGENTS_RULES.md.
Prepare a commit but wait for my approval.

๐ŸŽฏ Example Session

You:

I've downloaded and extracted TYPESCRIPT.zip in my project.
Please read AGENTS_RULES.md completely before we start.

AI:

I've read AGENTS_RULES.md. Key points understood:
- TDD workflow (Red-Green-Refactor)
- Write tests first, 90% coverage minimum
- TypeScript best practices (no any, explicit types, Result pattern)
- Git rules (no stash/push, ask approval)
- Conventional commits

Ready to start. What would you like to implement?

You:

Let's implement user authentication. Create a todo-list first.

AI:

Todo-list for user authentication:
1. [ ] Write acceptance test for user registration
2. [ ] Implement User domain model
3. [ ] Write unit tests for password hashing
4. [ ] Implement authentication service
5. [ ] Write integration tests
6. [ ] Verify 90% coverage

Shall I proceed?

๐Ÿ”ง Troubleshooting

AI Doesn't Follow Rules

Say:
You're not following the rules from AGENTS_RULES.md.
Please re-read the [SECTION] section and try again.

AI Commits Without Asking

Say:
STOP! According to AGENTS_RULES.md, you must NEVER commit without my approval.
Please revert and ask for approval first.

AI Skips Tests

Say:
You skipped writing tests first. According to AGENTS_RULES.md,
we follow TDD: write failing tests BEFORE implementation.
Let's start over with the test.

AI Uses Bad Practices

Say:
This code violates CODE_QUALITY_PRINCIPLES.md.
Please review the [PRINCIPLE] section and refactor.

๐Ÿ’ก Pro Tips


๐ŸŽ“ Summary

  1. Download the language pack (.zip)
  2. Extract to your project directory
  3. Tell AI to read AGENTS_RULES.md completely
  4. Verify AI understands the rules
  5. Start coding with TDD workflow
  6. Reference files during development

Remember: The AI agent must read AGENTS_RULES.md completely before starting any work. This ensures consistent, high-quality code following TDD and DDD principles. ๐ŸŽฏ