Survey Question Writer

What it does

Starts from a single stated research goal, for example why trial users don't upgrade, and writes 6-10 questions that actually test that goal, mixing multiple-choice, rating scale, and one open-ended question, while avoiding leading or double-barreled phrasing. Also suggests where each question type is best asked, in-app, email, or exit survey.

Install

This skill isn't packaged in an installable registry. It's a plain SKILL.md file you can copy directly. Copy it from the SKILL.md tab and save it at the path for your assistant:

Use it in your AI assistant

Claude Code

.claude/skills/survey-question-writer/ (or ~/.claude/skills/survey-question-writer/ for all projects)

GitHub Copilot (VS Code)

.github/skills/survey-question-writer/

OpenAI Codex

.codex/skills/survey-question-writer/

Cursor

.cursor/skills/survey-question-writer/ (or ~/.cursor/skills/survey-question-writer/ globally)

Compatibility

Any WordPress site; output is plain question text and question types for you to paste into whatever survey or form plugin you use.

The complete SKILL.md for this skill:

---
name: survey-question-writer
description: Use when a user wants survey questions drafted for a specific research goal, or asks to "write a survey" or "draft feedback questions".
---

# Survey Question Writer

## When to use
The user has a specific thing they want to learn from customers or users and needs a short, unbiased set of survey questions built around that goal, not a generic feedback form.

## Procedure
1. Restate the research goal in one sentence and confirm it's specific enough to write questions against.
2. Draft 6-10 questions directly tied to that goal, favoring closed questions for easy analysis, plus one open-ended question for texture.
3. Check every question for leading language and double-barreled phrasing (asking two things at once).
4. Order questions from easy and general to more specific, with demographic questions last.
5. Note a suggested survey length and best delivery point (post-purchase, exit-intent, email) for the batch.

## Guardrails
- Never write a question that presumes a positive answer or nudges toward one outcome, neutral phrasing only.
- Keep the survey short enough to plausibly finish in under 3 minutes, cut questions that don't map to the stated goal.