prompt-to-image
Prompt To Image
Generate prompt-based images with a bundled Node script that calls provider REST APIs directly.
This skill is for creating new images from prompts, and it also supports basic reference-image-guided generation for specific OpenAI and Gemini models. It is not for complex multi-step compositing workflows or rendering markdown into graphics.
Use This Skill For
- OpenAI image generation with current GPT Image models such as
gpt-image-1.5(default),gpt-image-1, orgpt-image-1-mini - Gemini image generation with current Google image models such as
gemini-2.5-flash-image(Nano Banana),gemini-3.1-flash-image-preview(Nano Banana 2), orgemini-3-pro-image-preview(Nano Banana Pro) - Reference-image prompting and image edits when the provider/model supports it
- Single images or small batches of prompt variations
- Tasks where the user wants a local image file as output
Do Not Use This Skill For
- Markdown, quote cards, code screenshots, or social post rendering from text blocks
Use
markdown-to-imageinstead. - Heavy multi-step image-edit workflows, advanced compositing, or provider-agnostic reference pipelines
More from ericjy/inkframe
markdown-to-image
Render markdown as beautiful styled images using the inkframe CLI. Use this skill whenever the user wants to turn text or markdown into a visual image, create social media graphics (tweets, Instagram posts, LinkedIn cards, quote cards, memes), generate code snippet images, make visual cards or posters from content, or produce any image from text/markdown. Also trigger when the user mentions 'inkframe', asks to 'render markdown', 'markdown to image', 'text to image', or wants to make content visually shareable. Even if they just say 'make this look nice as an image' or 'create a graphic for this' — use this skill.
24tier-list-image
Generate a recognizable tier-list image from rankings using a fixed HTML template and local browser rendering. Use this skill whenever the user wants a tier list, S/A/B/C ranking graphic, power ranking image, ranked comparison board, or says things like 'make a tier list', 'rank these visually', 'turn this into a tier-list image', or 'generate a shareable ranking graphic'. Trigger even when the user only gives rough ranking text and expects the agent to infer the structured tiers. Prefer agent-browser for local screenshots when available. Do not use this for generic posters, charts, or freeform markdown cards.
2