Skip to main content

How to Prompt Claude Well: 7 Basics of Writing Prompts

The same Claude gives very different results depending on how you write your prompt. Here are 7 core principles to get more accurate, useful answers.

🌐 This article was machine-translated and may contain inaccuracies. Read the Korean original if in doubt.

💬 1 copy-ready prompt in this article →

Even with the same Claude, the accuracy and usefulness of its answers change a lot depending on how you write your prompt. A good prompt is not a hard skill — it starts with a few principles. Master these seven and your results will improve noticeably.

🟢 Model references match the current lineup · model notice · Fable subscription
🟢 Model references match the current lineup · Claude Opus 5.5 / Claude Sonnet 5 / Claude Haiku 4.5 (higher tier: Claude Fable 5.1). This notice changes only when Anthropic ships a new model.

Fable 5 and 5.1 subscription (updated September 7, 2026): Claude Fable 5.1, released September 1, 2026, is the current Fable model and Fable 5 is now legacy. Plan terms are the same for both — Max and Team Premium plans include Fable at up to 50% of the weekly usage limit; Pro and Team Standard use usage credits (

🟢 Model references match the current lineup · Claude Opus 5.5 / Claude Sonnet 5 / Claude Haiku 4.5 (higher tier: Claude Fable 5.1). This notice changes only when Anthropic ships a new model.
0/M input, $50/M output tokens). The one-time
🟢 Model references match the current lineup · Claude Opus 5.5 / Claude Sonnet 5 / Claude Haiku 4.5 (higher tier: Claude Fable 5.1). This notice changes only when Anthropic ships a new model.
00 credit applied only to the Fable 5 transition and is not offered for 5.1. Some coding and debugging requests may be answered by an Opus model due to a security classifier (both models). See the Fable 5.1 guide and the Fable 5 availability guide for details.

⚡ TL;DR
  • Seven principles of good prompts — clarity · context · format · examples · steps · role · don'ts.
  • Not a hard skill — just following a few principles.
  • Clear & specific plus enough context makes the biggest difference.
  • Learn from bad vs good examples.
5 ingredients of a good prompt Role Context Task Examples Output format ↓ More accurate, useful answers

1. Be clear and specific

Instead of "summarize this," try "summarize the meeting notes below into 3 key decisions and action items per owner." The clearer you are about what and how, the more accurate the answer.

2. Provide enough context

Claude does not know your situation. Sharing the target reader, goal, field, and constraints leads to a much better fit — for example, "explain as if to a new hire with no marketing background."

3. Specify the output format

Decide length (e.g., 3 paragraphs), format (table, list, code), and tone (formal, friendly) up front to avoid rework. Try "as a table," "under 200 characters," or "as a numbered list."

4. Show examples

Showing one or two examples of the result you want (often called few-shot) helps Claude match the format and tone precisely. This is especially effective for repetitive, fixed-format tasks.

5. Break complex tasks into steps

For multi-step problems, asking Claude to "think step by step" improves accuracy. Splitting the process works better than asking for everything at once.

6. Assign a role

Setting a role (persona) such as "you are an experienced copy editor" aligns the answer's perspective and expertise with that role.

7. State what NOT to do

Spelling out what to avoid (e.g., "no jargon," "stick to confirmed facts, no guessing"), in addition to what you want, makes the result cleaner.

Vague vs specific request Vague "write a marketing post" → inconsistent results Specific audience · goal · format length · constraints → close in one try

Bad vs. good example

❌ Bad: "Write a marketing post."
✅ Good: "Write an Instagram post promoting an eco-friendly tumbler for people in their 20s, in a friendly tone, within 3 sentences, including 5 hashtags."

Packing in the audience, goal, format, length, and constraints like the second example gets you close to the result in a single try.

Wrap-up: a prompt is a conversation

It does not have to be perfect from the start. If the answer falls short, refine it with follow-ups like "shorter," "add an example," or "make it a table." The key is to ask specifically and improve iteratively.

Go deeper by topic

Each principle is covered in more depth here.

The principles in this article align with Anthropic official prompt engineering guidance (platform.claude.com/docs).

Frequently asked questions

How do I write good prompts?

Be clear and specific, give context, specify the output format, show examples, break complex tasks into steps, assign a role, and state what not to do.

What's the most important principle?

Writing clearly and specifically, plus giving enough context. Those two alone greatly improve answer quality.

Do examples really help?

Yes. Showing two or three examples of the format/style you want matches results more accurately than long explanations.

💬 Prompts for this problem

Copy, paste into Claude, and swap the parenthesised bits for your own situation.

  • 내 질문이 왜 원하는 답을 못 받았는지 봐줘

    Use when: 답변이 원하는 것과 자꾸 다르게 나올 때

    클로드한테 이렇게 물어봤는데 원하는 답이 안 나왔어. 뭐가 문제인지 봐줘.
    
    내 질문: "마케팅 아이디어 좀 줘" (여기에 네가 실제로 쓴 질문을 그대로)
    받은 답: 너무 뻔하고 일반적인 목록만 나왔어. (여기에 어떤 점이 아쉬웠는지)
    
    뭐가 빠져서 이렇게 됐는지 짚어주고, 같은 걸 물어보되 원하는 답이 나올 법한 질문으로 다시 써줘. 왜 그렇게 고쳤는지도 짧게.

All prompts →

Was this helpful?

Keep reading