jaytel0jaytel0Community·1 Agent Skills Included

taste

Turns reference images into reusable visual design rules

Converts curated reference images into a concrete, reusable SKILL.md file of visual design rules. Analyzes layout, spacing, color, typography, and hierarchy across images, then fuses findings into one consistent style system. Eliminates vague aesthetic guesses so generated interfaces follow exact, repeatable visual constraints.
npx skills add jaytel0/taste --all -g -y
Available:

Tells the agent how to run the local image-to-skill pipeline, which API keys to use, where generated skills are saved, and which credential and storage guardrails to follow.

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Taste?

Run `npx skills add jaytel0/taste --all -g -y` in your terminal to install the skill globally.

What does Taste do?

Taste analyzes a set of reference images and generates a SKILL.md file containing concrete visual design rules for typography, color, spacing, and layout.

How do I generate a design skill from my own images?

Place JPG, PNG, or WebP images in the reference-images folder, add your OpenAI and Anthropic API keys to .env.local, then run `npm run taste`. The finished skill is written to .taste/runs/<run-id>/SKILL.md.

Which models does Taste support?

It uses OpenAI and Anthropic vision models by default, and also supports OpenRouter or Vercel AI Gateway with a single API key.

Do I need design experience to use Taste?

No. You only supply reference images you like, and the pipeline extracts the visual rules and writes the skill for you automatically.

Related Repositories in Design & Creative

View All in Design & Creative