# How to Run the Prompt

1. **Select a model** that supports image input (e.g., `gpt-4-vision-preview`).
2. **Set the variables**:
   - `design_image_url` – URL to your screenshot (optional).
   - `design_description` – Brief description if no image.
   - `review_criteria` – Comma‑separated list (e.g., `accessibility, visual hierarchy, typography`).
3. **Paste the prompt** (the `body` field) into the model's chat window.
4. **Copy the markdown output** into your documentation or directly into a Figma comment block.

**Tip:** Store the prompt in a reusable snippet within your design tool plugin for one‑click access.
