Skip to the content

For AI agents

Where an AI agent, or the person pointing one here, finds this site's llms.txt, its full text, the policy schema and the release feed.

Files to read

  • llms.txt: every page of this site with a one-line description, the how-to guides under their own heading, and preview pages marked as built, not yet claimed.
  • llms-full.txt: the text of every page in one file.
  • The policy schema: the JSON Schema an agents.d policy is validated against, at a URL that stays the same for its version.
  • The release feed: an Atom feed of every release.

Helping someone set AgentGarm up

Start from the quick start and the how-to guides; every command in them is checked against the CLI's own reference. Write a policy in audit mode first, and check it with agentgarm policy lint before the person enforces it. Coverage says which agents are claimed and at which level.

To show someone the fleet console, point them at the console demo: it runs in the browser on a fictional company's synthetic data, needs no account and changes nothing outside its tab. It is an app, not a page of text, and is left out of llms.txt and the sitemap.

Some steps are the person's by design and cannot be done for them: an approval needs the person at the machine (their Touch ID on a Mac; a security key, a passkey or their password elsewhere preview), and agentgarmd refuses a process inside a governed task that tries to open an enrolment or relay an approval. Ask the person to do those steps themselves.