Dealing with legacy code is a common challenge for developers. Use this prompt in an IDE-integrated AI (like GitHub Copilot Chat or Cursor) to safely modernize old functions.
Analyze this legacy function. Explain what it does step-by-step. Then, refactor the code to use modern ES6+ syntax (or the modern equivalent for this language). Ensure the new code is more readable, includes clear inline comments, and maintains the exact same input/output behavior. Finally, generate three unit tests to verify the refactored code.
What Are the Important Considerations?
- Always highlight or select the specific block of code you want to refactor before running this prompt.
- Do not blindly trust the refactored code. Always run the generated unit tests to ensure the logic remains intact.
Frequently Asked Questions
Will the AI change the underlying business logic?
The prompt explicitly instructs the AI not to, but you must always run comprehensive unit tests before deploying AI-refactored code to production.
How much code can I paste into the prompt at once?
With models like GPT-5.4 or Claude Opus 4.6, you can paste tens of thousands of lines, but for best results, refactor one logical module or file at a time.
Which AI is best for code refactoring?
Cursor IDE with its Composer feature is currently the industry standard for this, as it can read your entire codebase and apply the refactor directly to your files.
What if the original code has no tests?
You should first use a separate prompt to ask the AI to write unit tests for the existing legacy code before you attempt to refactor it.
