From gpt-imagen
Generate or edit raster images for product design, frontend concepts, game assets, marketing visuals, and conceptual paper figures. Use when Claude should create a real image file through Codex or the OpenAI Images API instead of only suggesting prompts.
How this skill is triggered — by the user, by Claude, or both
Slash command
/gpt-imagen:gpt-imagenThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill when the user wants Claude to actually produce an image asset, not just brainstorm one.
Use this skill when the user wants Claude to actually produce an image asset, not just brainstorm one.
The gpt-imagen binary handles provider detection automatically. Do not check for API keys or Codex login status yourself. Always run the binary and let it decide. It will:
Never refuse to run the binary because you think a provider is missing. Never ask the user for an API key before trying. Just run it.
gpt-imagen binary to generate the image.--out.--image flags.Use this structure when the user prompt is vague:
Use case: <category>
Asset type: <where it will be used>
Primary request: <main ask>
Scene/backdrop: <setting>
Subject: <visual subject>
Style/medium: <photo | illustration | 3D | pixel art | infographic>
Composition/framing: <viewpoint and layout>
Lighting/mood: <lighting and tone>
Color palette: <palette>
Constraints: <must include or preserve>
Avoid: <negative constraints>
cat >/tmp/gpt-imagen-prompt.txt <<'EOF'
<structured prompt>
EOF
gpt-imagen --out <optional-path> - < /tmp/gpt-imagen-prompt.txt
When images are present:
gpt-imagen --image /abs/path/ref1.png --image /abs/path/ref2.png --out <optional-path> - < /tmp/gpt-imagen-prompt.txt
npx claudepluginhub aymenbouferroum/gpt-imagen --plugin gpt-imagenGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.