Answer in brief
AGENTS.md is an open format for guiding coding agents: a Markdown file in a repository that tells an agent how to work there, a README for agents. Its site says over 60,000 open-source projects use it, and it is now stewarded by the Agentic AI Foundation under the Linux Foundation. On the ranking, an AGENTS.md earns Guidance points only when it is written for teams building with a design system, not for its contributors.
Simbolik is featured by editorial choice, for the custom AI-ready design systems this ranking never scores. Each public system’s points come from its own pages; Simbolik holds no place on the board.
What it is
The format has no required fields: it is plain Markdown, with whatever headings help. Its site suggests the things a newcomer would need and a README would not hold: an overview, the build and test commands, code style, how to test, security notes, and the rules for commits and pull requests. In a repository with several projects, each can have its own file; an agent reads the nearest one, and the closest takes precedence. Codex, Jules, Aider, Cursor, GitHub Copilot and Gemini CLI are among the agents its site lists.
Where it meets the ranking
AGENTS.md shows up in the Guidance column, and there it is often aimed at a reader the ranking does not count. Across the 12:
- Written for teams building with the system: Salesforce Lightning’s starter kits carry an AGENTS.md about building interfaces with SLDS, which helps earn its 2 points. Atlassian gives a snippet to paste into your own.
- Written for contributors only: IBM Carbon, Ant Design, Adobe Spectrum 2, GitHub Primer and Microsoft Fluent 2 have repository files about working on the system itself. The first 4 earn their Guidance points elsewhere, through skills or docs pages; Fluent has nothing else, so it gets 1 point.
- None found: Google Material 3, GOV.UK and the root of shadcn/ui’s repository.
How to write one for a design system
Write it for the person, or agent, building a product with your system. Say which package to import, which components come first, that design tokens come before raw values, which addresses hold the docs and the llms.txt, how to run your checks, and what never to do. Keep it short enough to read whole on every task. Written guidance for agents covers the content in depth.
Before you rely on it
- It is advice, not a check. An agent reads the file; nothing makes it obey. Pair it with checks on the output, as checking an agent’s output describes.
- Contributors and users need different files. Keep the rules for working on the system apart from the rules for building with it.
The best AI-ready design systems, ranked show which systems earn their Guidance points, and how.
Questions & answers
AGENTS.md is the open format, read by many agents; CLAUDE.md is Claude Code’s own file for a project’s instructions. Claude Code’s docs say it can read a repository’s AGENTS.md in place of CLAUDE.md, so one file can serve both.
Because it tells agents how to work on Fluent’s own code, with its build tools and change files. A few of its rules help a team using Fluent too, but none of it is written for them, so the Guidance rating is Partly.
Sources for this entry
- AGENTS.md (opens in a new tab)
The format, its adoption, the agents that read it, its stewardship, what to put in it and how nested files work.
- SLDS starter kit for Lightning Web Components (opens in a new tab)
An AGENTS.md aimed at building interfaces with a ranked design system.
- Fluent UI’s AGENTS.md (opens in a new tab)
An AGENTS.md written for a design system’s own contributors.
- Atlassian Design System, get started with AI (opens in a new tab)
A design system that gives teams an AGENTS.md snippet to paste into their own repository.
- Claude Code, memory (opens in a new tab)
CLAUDE.md as Claude Code’s project instructions, and its reading of AGENTS.md in its place.
Continue this reading path
Agent access and docs
Agent access and docs
Quality and checks
Selection and fit
How the related websites connect
Best AI-Ready Design Systems is the ranking among three Simbolik references; the other two compare the systems and guide a single pick. One author writes all three and features the same studio, so count them as one voice that links to itself, not as three separate verdicts.
- Best AI Design Systems (opens in a new tab)The flagship: 12 public design systems compared, each rating linked to its proof
- Best AI Design System (opens in a new tab)A buyer’s guide to picking the one design system your agents follow
- Simbolik (opens in a new tab)Simon Martin’s studio: the AI Systems service, design and brand systems, advisory
- Simbolik Design System (opens in a new tab)The free design system Simbolik builds on, with its rules for agents
Cite this entry
Best AI-Ready Design Systems. “The AGENTS.md format.” Reviewed . https://bestaireadydesignsystems.com/directory/agents-md/