Hands-OnAI
← All skills
Image & videoNewFreeWorks in Claude Code · Claude app · Codex

GPT Image 2 Style Library

Turn a rough image idea into a production-ready GPT Image 2 prompt.

Part of awesome-gpt-image-2 by @freestylefly ↗

GPT Image 2 Style Library gives Claude a catalog of proven image templates — product shots, posters, UI screens, infographics, brand visuals, characters and more — with style tags, scene tags and pitfalls for each. Describe what you want and it picks the best-fitting template and writes a copyable prompt with layout, style, text, aspect ratio and constraints. It writes the prompt; you run it in ChatGPT or any tool that uses GPT Image 2.

What it does

  • Matches your request to a template category, visual style and scene from the library.
  • Offers 2–3 options with short reasons when more than one template fits.
  • Builds the prompt in blocks: subject, composition, style and materials, text, aspect ratio, constraints.
  • Answers in your language (English or Chinese examples in the library).

When it kicks in

Choose GPT-Image2 / gpt-image-2 visual styles and industrial prompt templates from the awesome-gpt-image-2 style library. Use when an agent needs to create, rewrite, classify, or improve image-generation prompts with repository-backed templates, categories, style tags, scene tags, pitfalls, and example cases.

Works with

  • Your image idea
  • Brand or product details
  • Text that must appear

Topics the skill covers

  • Example Output
  • Workflow
  • Output Defaults
  • Maintenance

What's inside the download

  • SKILL.md — the prompt-building workflow.
  • references/style-library.md — the templates, style tags, scene tags and example cases.
  • assets/ — one example output image.
  • LICENSE — MIT.

5 files in the .zip:

  • Skill InstructionsSKILL.md2.4 KB
  • LicenseLICENSE1 KB
  • OpenAIagents/openai.yaml211 B
  • City Life System Mapassets/city-life-system-map.png1.9 MB
  • Style Libraryreferences/style-library.md25.7 KB

How to install (2 minutes)

Same download for Claude Code, Claude app, Codex — follow the step for the tool you use.

  1. 1Download the .zip and unzip it — you’ll get a folder called gpt-image-2-style-library.
  2. 2In the Claude app: open Settings → Capabilities → Skills, choose Add / Upload, and select the folder (or the .zip).
  3. 3In Claude Code: drop the folder into your project’s .claude/skills/ directory, then type /skills to confirm it loaded.
  4. 4In Codex: drop the folder into ~/.agents/skills/ (or your project’s .agents/skills/), then type /skills — or call it directly with $gpt-image-2-style-library.
  5. 5Then ask for an image prompt — for example “use the GPT Image 2 style library to write a poster prompt for my workshop” — and paste the result into your image tool.

New to skills? Our Claude Cowork beginner's guide shows how they work inside Claude.

Want to get the most out of your new skill?

Inside Hands-On AI Club we show you exactly how to use skills like this in real work — with step-by-step courses and a community to help you.