My last project is now live: AI Intent Driven Development (IDD) guidelines and instructions for AI Coding Agents, AI Coding Assistants, and LLMs.
What a heck is an Intent, you may ask to yourself?
An Intent is a self-contained document that describes a user request. It is composed of three main sections: WHY (the motivation), WHAT (the requirements, often in Gherkin language), and HOW (a detailed, step-by-step implementation plan defined with tasks). This approach ensures clarity and alignment before any code is written.
This project delivers this guidance as a collection of markdown files that instruct AI agents to follow specific guidelines for architecture, development workflow, and more. It supports both language-agnostic and language-specific guidance. Currently, the project focuses on Elixir and Phoenix, with plans to support other BEAM languages and frameworks in the near future. Other programming languages and frameworks will also be supported.
Github repository: https://github.com/Exadra37/ai-intent-driven-development/
Intent Specification: https://github.com/Exadra37/ai-intent-driven-development/blob/main/INTENT_SPECIFICATION.md
Intent Example: https://github.com/Exadra37/ai-intent-driven-development/blob/main/INTENT_EXAMPLE.md
I would love your feedback on:
- The clarity and usefulness of the Intent specifications and example.
- Ideas for expanding to other languages or frameworks.
- Where you see this approach being helpful (or not).
- On how practical you find this guidelines and workflow for guiding AI coding agents in a BEAM environment (Elixir and Phoenix).
