Skip to content
PromptsBuddy
All prompts
FundamentalsbeginnerStaff pickClaudeChatGPTGeminiGrokPerplexity

Make it ask before it answers

Four lines that stop a model guessing at what you meant. It interviews you until it can actually do the task, which turns a mediocre first answer into a usable one.

RonyRony · Founder, PromptsBuddyAug 25, 2026v1.0
The prompt· 0/1 variables filled
~103 tokens
Before answering, ask me clarifying questions until you are confident you can
do this correctly. Ask them one or two at a time, not as a list of ten.

Stop asking once more questions would not change your answer, then say so and
proceed.

Do not guess at anything I have not told you. Do not fill gaps with plausible
defaults. If you use an assumption anyway, label it inline as an assumption.

THE TASK:
{{TASK}}

Fill in the blanks

Run it in:ClaudeChatGPTGemini

What goes in each blank

TASK
What you want done, stated however roughly you like.e.g. Write the launch plan for our new pricing tier.

Notes from people who use it

  • This is the single highest-leverage prompt in the library, and it's four lines. Most bad output is a guessing problem, not a model problem.
  • 'One or two at a time' matters — a list of ten questions gets skimmed and half-answered.
  • If it asks nothing and proceeds, your task description was genuinely complete. That's worth knowing too.

Use this as a skill

A prompt is a skill with the serial numbers filed off. Install it once and your agent reaches for it on its own.

Paste this into ChatGPT, Claude Code, Codex, Cursor, or whatever you use:

Install the make-it-ask-before-it-answers skill globally from https://promptsbuddy.com/prompts/make-it-ask-before-it-answers/skill.md
Add to Cursor

.claude/skills/make-it-ask-before-it-answers/SKILL.md

---
name: make-it-ask-before-it-answers
description: "Four lines that stop a model guessing at what you meant. It interviews you until it can actually do the task, which turns a mediocre first answer into a usable one. Use when the user asks for help with fundamentals tasks like prompting, accuracy, short."
license: CC-BY-4.0
metadata:
  source: https://promptsbuddy.com/prompts/make-it-ask-before-it-answers
  author: "Rony"
  version: "1.0"
---

# Make it ask before it answers

Four lines that stop a model guessing at what you meant. It interviews you until it can actually do the task, which turns a mediocre first answer into a usable one.

## Inputs to collect first

Ask the user for anything below that they have not already given you. Do not
invent values for these.

- `TASK` — What you want done, stated however roughly you like. (e.g. Write the launch plan for our new pricing tier.)

## Instructions

Before answering, ask me clarifying questions until you are confident you can
do this correctly. Ask them one or two at a time, not as a list of ten.

Stop asking once more questions would not change your answer, then say so and
proceed.

Do not guess at anything I have not told you. Do not fill gaps with plausible
defaults. If you use an assumption anyway, label it inline as an assumption.

THE TASK:
{{TASK}}

## Notes from people who use this

- This is the single highest-leverage prompt in the library, and it's four lines. Most bad output is a guessing problem, not a model problem.
- 'One or two at a time' matters — a list of ten questions gets skimmed and half-answered.
- If it asks nothing and proceeds, your task description was genuinely complete. That's worth knowing too.

---

Make it ask before it answers · by Rony · v1.0
From PromptsBuddy — https://promptsbuddy.com/prompts/make-it-ask-before-it-answers
Licensed CC BY 4.0.

Every prompt is also available at /prompts/make-it-ask-before-it-answers/skill.md — see all install options.

Why this outperforms longer prompts

The most-shared prompts in circulation are short ones that fix a recurring failure, and this fixes the most common one: the model does not have enough information, so it invents the missing half and answers fluently anyway.

You cannot spot that in the output. Fluent guesses and fluent knowledge look identical. Moving the guessing to the front — as questions — makes it visible.

The stopping condition

Without "stop asking once more questions would not change your answer", the interview never ends and you spend more effort answering questions than you would have spent writing the brief. The instruction to name assumptions inline is the safety net for whatever it decides not to ask about.

#prompting#accuracy#short#fundamentals

Try these next

Fundamentalsbeginner

Kill the flattery

A standing instruction that removes the praise, the restating, and the closing summary — so the reply is only the part you actually needed.

ClaudeChatGPTGemini1 vars
Lena FischerLena Fischer