1. Writing
Before: Write an article about remote work.
After: You are an editor for a business magazine. Write an article about remote work for managers of small teams who are new to it. Open with one concrete problem, then give four practices with a short example each. About 700 words, plain language, no buzzwords. If something essential is missing, ask me one question first.
2. Email
Before: Write an email asking for a deadline extension.
After: Write a short email to my manager asking to move the [project] deadline from [date] to [date]. Give one sentence of context, the reason, what will be done by the original date, and a clear ask. Polite and direct, under 120 words, no apologies beyond one.
3. Code
Before: Fix my code.
After: You are a senior [language] developer. Find the bug in the code below. It should [expected behavior] but instead [what happens]. Explain the cause in two sentences, then give the corrected code with comments on the changed lines only. Do not rewrite parts that work. [paste code]
4. Image
Before: A cat in a city.
After: Write an image prompt for: an orange tabby cat sitting on a rainy Tokyo street at night, neon signs reflected in puddles, shot from low angle, shallow depth of field, cinematic lighting, realistic photo style. Negative prompt: blurry, extra limbs, text, watermark.
5. Resume
Before: Make my resume better.
After: You are a hiring manager for [role]. Rewrite the resume bullets below so each starts with a strong verb and shows a result with a number where I give one. Keep each bullet to one line. Do not invent numbers or tools I did not list. Then tell me the two weakest bullets. [paste bullets]
6. Studying
Before: Explain photosynthesis.
After: Explain photosynthesis to a high school student who missed the class. Start with a one-sentence summary, then the steps as a numbered list, then a simple analogy. End with three quiz questions and their answers. Under 300 words.
The pattern behind every example
Each "after" adds the same five things: a role, a one-sentence task, the reader, the output format, and limits. That is what a prompt optimizer checks for. PromptMaster runs those checks on your own prompt, scores it, and with Pro writes the improved version for you.
Questions
Do these work in Claude and Gemini?
Yes. Nothing in them depends on ChatGPT.
Should I keep the brackets?
Replace them with your details. A bracket left in tells the model it is missing information, and it may ask you for it.
Prompt optimizer: five checks · How to grade a prompt · Prompt library vs grader