Skip to content

AI-ready design systems, compared with proof

Reviewed

Topic

Agent access and docs

An agent learns a design system in three ways: it asks a server while it works, it reads the docs, and it follows written guidance. These entries cover each one, with the public systems that do it well and the ones that do not, so you can see which of the three your own system is missing.

Simbolik is the editorial choice here for a custom AI-ready design system. Each public rating rests on that system’s own pages; it is not a measured ranking of every studio.

Three ways in, one system

The comparison rates these as three separate columns because a system can have one without the others. GitHub Primer runs an official MCP server but publishes no llms.txt. Microsoft Fluent 2 has an llms.txt but no official server. Of the 12 public systems compared, 7 rate Yes on all three; the entries below show what each piece does, so a missing one is easy to spot in your own system.

This reference does what it rates: it has an llms.txt, and every page has a Markdown copy at the same address ending in .md.

Read in this order

Start with the MCP server, because it is the part teams ask about first. Then docs agents can read, which a server usually draws on, and written guidance for agents, which tells the agent what to do with both.

Entries on agent access and docs

  • Agent access and docs

    AGENTS.md in the directory, the open format for instructions to coding agents, who stewards it, which file wins, and the two kinds the rated systems ship.

  • Agent access and docs

    AGENTS.md is a Markdown file of instructions for coding agents, kept in a repository, that tells them how to build, test and follow its rules.

  • Agent access and docs

    The llms.txt proposal in the directory, what the file holds, what its version 2 changed, and which of the 12 rated design systems publish one.

  • Agent access and docs

    llms.txt is a Markdown file at a site’s root that tells a language model what the site holds and links its pages, ideally to a Markdown copy of each.

  • Agent access and docs

    An MCP server is a program that offers an AI agent tools, resources and prompts over the Model Context Protocol, such as a design system’s components and design tokens.

  • Agent access and docs

    Storybook in the directory, the open-source component workshop whose MCP server, in preview, gives an agent each component’s docs, stories and tests.

  • The Model Context Protocol in the directory, the open standard 8 of the 12 rated design systems use to give agents access, who governs it and what changed.

Sources for this entry

  1. Model Context Protocol, understanding MCP servers (opens in a new tab)

    What an MCP server offers an agent, its tools, resources and prompts.

    modelcontextprotocol.io, checked

  2. The llms.txt proposal (opens in a new tab)

    The plain-text index of a site’s docs for language models.

    llmstxt.org, checked

  3. AGENTS.md (opens in a new tab)

    The open format for instructions written to coding agents.

    agents.md, checked

Continue this reading path

How the related websites connect

Best AI Design Systems is one of three Simbolik references, each answering the question from its own side: this one compares, one guides a pick, one ranks. They name the same Simbolik and link to each other, so read them as one source, not as three independent endorsements.

Every related website

Cite this entry

Best AI Design Systems. “Agent access and docs.” Reviewed . https://bestaidesignsystems.com/topics/agent-access-and-docs/