All products

Write image prompts with AI

Turn a still-image brief into a specific production prompt, separating what an edit must preserve from the changes it should make and the failures to avoid.

Write image prompts with AI by describing the picture rather than piling up quality adjectives. The image-prompt-writer skill organises a brief around a visible subject, composition, materials, lighting and purposeful constraints. It can prepare a new image request or a carefully bounded edit.

The deliverable is text. No generation provider, private workflow or API key is required by the skill itself. Reference-based work still needs an accessible, authorised reference, and the eventual image must be checked in the tool you choose.

What it does

  • Prepare a product-image generation brief
  • Separate preservation rules from edit instructions
  • Describe intentional lettering and annotations
  • Plan multiple views on a design sheet
  • Define negative space for a cover layout
  • Keep a reusable prompt independent of one provider

Install and use

  1. Supply the whole folder and references to a file-capable agent. Add authorised images only when needed; viewing them requires a compatible image-capable host. Prompt writing alone does not submit a generation job.
  2. Try: “Write a poster-image prompt for a rainy tram stop. Leave the upper third clear for our supplied headline; keep the vehicle visible and avoid invented branding.”
  3. Classify the task, fill concrete visual details and state visible-text intent. For edits, use separate Preserve, Edit and Avoid sections rather than scattering constraints through prose.
  4. Review the prompt against the brief. Request platform-specific formatting only for a known target, then inspect the actual generated result separately.

FAQ

Source

Made by JinUnique. Original skill: image-prompt-writer, published on GitHub under the MIT licence.

View on GitHub