Back to directory

color-expert

color-expert is an agent skill that turns a coding agent into a color-science reference, covering color spaces, accessibility contrast standards such as APCA and WCAG, palette generation, pigment mixing and historical color theory, backed by a large collection of source material.

color-expert

color-expert grew out of one skill file of color-theory notes into a three-layer knowledge base: a short SKILL.md of key facts and tool recommendations that loads immediately, an index file that helps the agent find the right reference for a specific topic, and a large set of deeper reference files with full video transcripts, article summaries, library documentation and research notes.

New material is added by pasting a link into the skill's own workflow: videos are transcribed and summarized, PDFs and documents are converted to Markdown, GitHub repository documentation is fetched, articles are extracted, books mentioned in videos are searched on the Internet Archive, and whole reference websites are scraped chapter by chapter, with tools and links collected into their own table.

A lightweight evaluation folder of trigger and task prompts lets the skill be reviewed against realistic usage rather than edited by feel alone. It is released under a Creative Commons Attribution license.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.