visual-qna

Installation
SKILL.md

Visual Q&A

Overview

Render a set of structured questions as an interactive form in a Playwright-driven browser, let the user click their answers (single- or multi-select, with an "Other" choice and a free-text note per question), then read the answers back as structured data. Good for batches of decisions where the native one-question-at-a-time terminal flow is slow.

Core idea: inject a reusable form engine into about:blank via the Playwright MCP — no server, no files, no ports.

When to use

  • A batch (≈3+) of multiple-choice questions: brainstorm decisions, option-picking, questionnaires, A/B/C choices.
  • You want per-question notes and an "Other" escape hatch alongside fixed options.
  • You want the answers handed back as structured data to act on.

When NOT to use

  • Mockups, wireframes, diagrams, visual design comparisons → that is a visual companion's job, not this. This skill is for answering questions, not looking at pictures.
  • One or two simple questions → just ask in the terminal, or use the native AskUserQuestion tool.
  • Playwright MCP is unavailable → see Requirements.
Installs
64
GitHub Stars
1
First Seen
May 27, 2026
visual-qna — mmcardle/ai-skills