This teaches you how to call Gemini and GPT image APIs from Claude Code with the right model for each job. Gemini handles photorealistic scenes and stock imagery better, GPT Image 2 is the only one that renders readable text (posters, OG images, infographics), and GPT 1.5 does transparent icons since GPT 2 can't. It walks through a five-part prompting framework (image type, subject, environment, technical specs, constraints) and includes API calling patterns for both providers. Useful when you're building a site and need hero images, service photos, or social cards without leaving your editor. The model selection matrix is practical and the text rendering distinction matters more than you'd think.
npx skills add https://github.com/jezweb/claude-skills --skill ai-image-generator