generating-images

Installation
SKILL.md

Generating images

Turn a description of what an image must achieve into one generated image, with a record of how it was made and alt text that discloses it is AI-generated.

Script paths below are relative to this skill's folder. Run each as one command with the folder's absolute path in front, never chained with &&, $(…), or heredocs. The script uses Node's built-in fetch; there is nothing to install.

Before generating: provider and cost

The image provider is chosen like this: --provider gemini or --provider openai when the person asked for one; otherwise whichever of GEMINI_API_KEY or OPENAI_API_KEY is set, Gemini first. If neither is set, ask the person to set one; never ask them to paste a key into the conversation.

Generating costs money, so the person approves first:

  1. Run node scripts/generating-images.mjs plan --count <n> for all the images planned.
  2. Show the person the provider, model, image count, and pricing link it prints, and wait for approval. When the images are part of a larger plan (for example a slide deck's visuals plan), include this in that approval instead of asking twice.
  3. Only then generate, passing --approved.

Generating one image

Installs
1
Repository
ken-guru/skills
GitHub Stars
1
First Seen
1 day ago