Best AI Prompts for Debugging Code (With Real Examples)

After debugging hundreds of issues with AI assistance, I’ve found that HOW you share the bug dramatically affects the quality of help you get. Here are my battle-tested debugging prompts.

For runtime errors:
“I’m getting this error: [paste error]. Here’s the relevant code: [paste code]. Here’s what I expected to happen: [description]. Here’s what actually happens: [description]. I’m using [language version] with [relevant libraries/frameworks].”

For logic bugs:
“This function should [expected behavior] but instead it [actual behavior]. Here’s the function: [code]. Here’s a test case that fails: [input → expected output → actual output]. Walk through the code step by step with this input and find where the logic diverges.”

For performance issues:
“This code runs in [current time] but I need it under [target time]. Here’s the code: [paste]. The data characteristics are: [size, shape, distribution]. Suggest optimizations ranked by likely impact.”

The most common mistake I see: People paste the error message and say “fix this” without providing context. You’ll get generic suggestions. Always provide the context. The more context, the better the fix.

What debugging prompt patterns work for you?