Debug an error properly

Forces the model to rank causes by likelihood before proposing a fix.

Works with ChatGPT, Claude, Gemini

Use it when: Getting unstuck on a stubborn bug.

Fill in the blanks

StackExpectedActualErrorCode

Your prompt

Act as a debugging partner.

Stack: [Stack]
What I expected: [Expected]
What happens instead: [Actual]
Error output:
[Error]
Relevant code:
[Code]

Return:
1. The three most likely causes, ranked, each with the evidence in my output that supports it.
2. One quick check per cause to confirm or rule it out.
3. Only after that, the fix for the most likely cause, with the changed code.

Do not guess at code I have not shown you. Ask for it instead.

Make it yours

Improve your own version, or save it with your usual details filled in.

Improve with AI Save to my vault

Get the extension Learn to write your own

More coding prompts

CodingChatGPT, Claude, Gemini
Build and explain a regex
A pattern you can actually read, with test cases.
Fill it in and copy
CodingChatGPT, Claude, Gemini
Document an API endpoint
Reference docs a stranger could implement against.
Fill it in and copy
CodingChatGPT, Claude, Gemini
Explain this code line by line
A walkthrough pitched at your level, plus the parts most likely to bite you.
Fill it in and copy
CodingChatGPT, Claude
Refactor without changing behaviour
Small, safe, reviewable steps instead of a rewrite.
Fill it in and copy

Fill in the blanks

Your prompt

Make it yours

More coding prompts

Build and explain a regex

Document an API endpoint

Explain this code line by line

Refactor without changing behaviour