testing

AAID : Augmented AI Development

Professional TDD for AI-Augmented Software Development

🔻

📋 Repository Purpose: This repository contains support resources for developers who want to follow the AAID workflow. The main methodology is introduced in our dev.to article (also available in the repo here) with detailed resources, templates, and tools provided here.

What is AAID?

AUGMENTED AI DEVELOPMENT AAID (/eɪd/ - pronounced like "aid") is a disciplined approach where developers augment their capabilities by integrating with AI, while maintaining full architectural control. You direct the agent to generate tests and implementation code, reviewing every line and ensuring alignment with business requirements.

You're not being replaced. You're being augmented.

This separates professional software development from "vibe coding." While vibe coders blindly accept AI output and ship buggy, untested code they can't understand, AAID practitioners use TDD (Test-Driven Development) to ensure reliable agentic assistance.

The AAID Advantage

• Predictable Development: TDD discipline keeps the AI under control - your control
• Full Understanding: You review and understand every line of code
• Production Quality: Tests ensure code meets business and technical requirements
• Comprehensive Testing: Unit tests (TDD) and acceptance tests (ATDD/BDD) ensure every aspect of the application is fully tested
• Speed + Stability: AI accelerates development while TDD ensures reliability

Built on proven foundations from Kent Beck's TDD, Dave Farley's Continuous Delivery, Acceptance Testing and Four-Layer Model, Robert C. Martin's Three Laws of TDD, and other battle-tested practices.

Quick Links

• 🚀 Start Here: Onboarding Guide - Gentle introduction to AAID (~60 minutes, 7 progressive levels)
• Complete AAID Guide - The full methodology and workflow
• Complete AAID Acceptance Testing Guide - Using Dave Farley's Four-Layer Model for ATDD/BDD
  • Acceptance Testing Workflow Diagram - Visual representation of the AAID Acceptance Testing process
• Demo Repository - TicTacToe CLI developed using the AAID workflow, demonstrating the Four-Layer Model, test isolation patterns, and hexagonal architecture
• Workflow Diagram - Visual representation of the AAID process
• Claude Code Plugins - Install aaid-tdd and aaid-bdd as Claude Code plugins that enforce the workflow

Repository Structure

├── docs/
│   ├── onboarding/       # Gentle introduction to AAID
│   └── aidd-workflow.md  # Complete AAID methodology guide
├── appendices/           # Supporting documentation
├── .cursor/commands/     # Cursor slash commands (copy to other tools' custom-command folders)
│   ├── planning/
│   ├── tdd/
│   ├── investigation/
│   └── misc/
├── rules/                # AI workflow rules (drop into .cursor/rules/ or CLAUDE.md)
└── plugins/              # Claude Code plugins (aaid-tdd, aaid-bdd)

Core Components

🚀 Onboarding

• docs/onboarding/guide.md: Gentle introduction to AAID with 7 progressive levels (~60 minutes total)
  • Start here if you're new to AAID or TDD

📙 Main Guide

• docs/aidd-workflow.md: The complete AAID methodology with detailed explanations, examples, and step-by-step workflows

📚 Appendices

• Appendix A: Acceptance Testing with Four-Layer Model
  • Complete workflow for ATDD/BDD using Dave Farley's architecture
• Appendix B: Reusable prompts and AI commands
  • Pre-written AI commands organized by purpose to speed up development
• Appendix C: AI workflow rules for your IDE
  • Ready-to-use AI agent instructions that enforce the AAID TDD workflow discipline
• Appendix D: Handling technical implementation details
  • Understanding the three Implementation Categories and how to apply AAID to technical implementation details
• Appendix E: Dependencies and Mocking
  • Understanding the four Dependency Categories and how to handle them in different test types

⚡ Reusable Prompts