While learning to code, I kept retyping the same kinds of prompts: explain this code, help me debug this, quiz me on this topic. So I started saving the ones that actually worked and turned them into an open-source collection called GPT Tips & Tricks.
What's inside
There are 7 categories, about 33 prompts in total:
- Coding: explaining code, reviewing it, writing tests, refactoring
- Debugging: finding bugs, decoding error messages, wrong output, slow code
- Study: learning topics, quizzes, study plans, summarizing notes
- Project: generating ideas, scoping, breaking work into tasks
- Documentation: READMEs, commit messages, code comments
- Career: mock interviews, resumes, explaining your projects
- Dev tools: Git, terminal, SQL, regex, environment setup
Example
Here's one from the debugging category:
I am a beginner programmer. Explain this error message in simple terms.
Tell me:
1. What the error means
2. What usually causes it
3. How I can fix it
4. How I can prevent it in the future
Error:
[PASTE ERROR HERE]
Asking for four separate parts forces GPT to teach the concept instead of just handing you a fix.
See real outputs
The repo has an examples/ folder with real GPT responses, so you can judge a prompt before using it. Every prompt is copy-paste ready, and you just fill in the [BRACKETS].
Repo
https://github.com/Abex04/gpt-tips-and-tricks
It's free and MIT licensed. If you have prompts that work well for you, pull requests are welcome.
Top comments (1)
Including real outputs gives beginners something concrete to inspect. For the study category, a quiz example that waits for an answer before giving feedback would complement the explanation prompts. Could the examples also include one weak output and why it needs checking?