auth.md
A machine-readable authorization and identity declaration convention that allows AI agents to autonomously provision restricted resources without requiring a human account, while supporting subsequent claiming by a human account. The mechanism implements an agent-friendly, progressive-trust provisioning model.
Timeline (last 90 days)
auth.md was proposed by WorkOS and is still in an early adoption stage; it lets services use a Markdown file to describe registration, authentication, and available permissions, supporting two modes: Agent verified and User claimed.
auth.md overlaps in functionality with OAuth Dynamic Client Registration (RFC 7591) and domain proposals such as Agent identity verification, and whether it can ultimately become a de facto standard remains uncertain.
Through the auth.md mechanism, Neon allows an AI Agent to provision a restricted database project before a human formally registers, which can later be transferred to the user who wants to keep it.
auth.md is essentially a machine-readable authorization and identity declaration convention.
auth.md draws inspiration from the concept of .md/.txt metadata files such as robots.txt or llms.txt.
All Facts (5)
auth.md was proposed by WorkOS and is still in an early adoption stage; it lets services use a Markdown file to describe registration, authentication, and available permissions, supporting two modes: Agent verified and User claimed.
50%Unverifiedauth.md overlaps in functionality with OAuth Dynamic Client Registration (RFC 7591) and domain proposals such as Agent identity verification, and whether it can ultimately become a de facto standard remains uncertain.
50%UnverifiedThrough the auth.md mechanism, Neon allows an AI Agent to provision a restricted database project before a human formally registers, which can later be transferred to the user who wants to keep it.
50%Unverifiedauth.md is essentially a machine-readable authorization and identity declaration convention.
50%Unverifiedauth.md draws inspiration from the concept of .md/.txt metadata files such as robots.txt or llms.txt.
50%