Description

The project aims to explore how to design effective system messages and prompts tailored for AI-driven code refactoring tasks. By refining these prompts, we can guide AI models to better understand code structure, identify improvement opportunities, and execute refactorings with minimal manual intervention. The project will delve into creating system messages that balance technical detail with flexibility, enabling AI to perform refactorings that maintain code functionality while improving readability, efficiency, and maintainability.

Goals

1. Develop Effective Prompts: Identify and test prompt structures that guide AI in recognizing refactoring patterns, best practices, and coding standards.
2. Evaluate Refactoring Quality: Define metrics to assess refactoring quality, such as code readability, maintainability, and performance improvements.
3. Automate Refactoring Workflow: Create a repeatable process that integrates these system messages with a chosen AI tool for code refactoring.
4. Document Findings and Best Practices: Compile a set of guidelines for designing prompts and system messages, applicable across different codebases and AI platforms. Resources

Resources

• AI Tools and APIs: GitHub Copilot: For in-editor code suggestions and refactoring insights. ChatGPT: To generate, evaluate, and refine code refactoring prompts and responses, leveraging conversational context.
• Sample Code Repositories