Prompting Notes

This page is a compact reminder for everyday prompting. It is meant for practical use: getting cleaner answers, improving code comments, and turning vague requests into structured tasks.

A Simple Prompt Structure

Use four basic parts when the model output is unstable:

1. Role: tell the model who it should act as.
2. Task: describe the concrete job to finish.
3. Constraints: define output format, length, style, or scope.
4. Examples: provide one or two examples when the format matters.

Typical Use Cases

• Ask the model to write comments for an existing code block.
• Ask for input, output, and behavior summaries before a function definition.
• Ask for a cleaner rewrite of notes, reports, or README text.
• Ask for a draft implementation first, then refine it in later rounds.

A Better Way To Iterate

• Start with a plain request.
• Add structure if the result is too vague.
• Add constraints if the result is too long or off-topic.
• Add examples if the format is still inconsistent.
• Keep only the useful context instead of pasting everything.

Good Habits

• Be explicit about the expected output.
• Split large tasks into smaller steps.
• Verify code and facts instead of trusting the first answer.
• Avoid sending sensitive keys, credentials, or private data.

Quick Template

1
2
3
4

Role: You are a senior Python developer.
Task: Explain this function and add a short comment block above it.
Constraints: Keep the explanation under 80 words.
Output: First the comment block, then a one-paragraph explanation.