File Conversion
Security scan passedConverts images, audio, video, documents and data with the tools on your machine, keeps your originals and checks every result.
Loading listing…
GPTImage2Skill
Security scan passedGenerate and edit images with OpenAI's GPT Image models from your agent, with a gallery of tested prompts and an image-to-prompt skill.
Pack · 2 skillsby wuyoscar· MediaUnclaimed · Yours? Claim it
Preview
4 images
What GPTImage2Skill does
About
GPTImage2Skill is a community project by wuyoscar: two Agent Skills and a command-line tool for OpenAI's GPT Image 2 and 2.5 models, plus a prompt gallery of more than 160 entries across 31 categories, from posters and infographics to product shots, pixel art and research-paper figures. What its skills teach your agent - gpt-image: generates or edits images through its packaged CLI, using your OpenAI API key. It asks which model to use when that's unclear, then passes it explicitly. - get-prompt-from-image: studies a reference image and writes a detailed prompt that recreates its look, ready to hand to gpt-image or another generator. What you can do with it - Turn a text prompt into an image, with size, quality, format and transparent-background options - Edit or restyle an image from one or more reference images, or inpaint the area a mask marks - Recreate any gallery entry by asking for it in plain words - Extract a prompt from a photo, illustration, logo or poster you like How to install it - In Claude Code: /plugin marketplace add wuyoscar/gpt_image_2_skill, then /plugin install gpt-image@wuyoscar-skills. - Or npx skills add wuyoscar/gpt_image_2_skill with the skills you want, Codex's built-in skill installer, or a manual copy. - The CLI runs with uvx from the repo and needs Python 3.11 or newer. Good to know - Image calls go to OpenAI's API with your OPENAI_API_KEY and may be charged by OpenAI; higher quality settings cost more and take longer. - It's a community project, MIT-licensed, not made by OpenAI. - Outputs are PNG, JPEG or WebP images; editable SVG or slide shapes need a separate step.
More from Media
Skills people install alongside GPTImage2Skill.
