Test Driven Development

This section outlines the process for doing test driven development.

Workflow

Prerequisites

1. Clear Requirements: Detailed feature requirements with defined scope, as outlined in Generating Requirements
2. Rules and instructions for AI: Set up project rules and instructions for AI for consistent and repeatable standards, patterns and conventions across your codebase
3. Capable Code Generation Model: Ensure you’re using the most capable LLM model for the task you are running to get good quality results

1. Create a new git branch

• Start with a new git branch dedicated to the feature.

2. Prompt the Coding Assistant

• Use an AI Coding Assistant (AICA) in Agent mode
• Use the three prompts in prompt-test-driven-development to do RED GREEN REFACTOR.
• First run the prompt to do step 1. RED: Write failing tests.
  • Reference your requirements files directly using the Coding Assistant Agent.
  • Review and refine the tests it generates
• Then run the prompt to do step 2. GREEN: Minimal code to pass.
  • Review and refine the code generated
• Finally run the prompt to do step 3. REFACTOR: Improve while keeping tests green
  • Review and refine the refactored code generated

Follow-up tasks

• Create merge request for peer review
• Update documentation using prompt-add-update-documentation
• Update AI rules and prompt library based on lessons learned

Guidelines

Use an apporpriate model: Ensure you are using a suitable model to implement this effecively.

Get consistent results: As per the four pillars, use the combination of rules and instructions for AI, prompts and clear product requirements.

Review everything generated: You must review and refine everything generated by the agent.

Next -> Testing