Agent Definition Language (ADL) is a new open standard that provides a declarative, framework-agnostic specification for defining AI agents. Similar to how OpenAPI standardized API contracts, ADL consolidates agent definitions into structured documents that include role, capabilities, tool permissions, workflow logic, safety constraints, and governance metadata. This addresses current fragmentation where agent behavior is scattered across prompts, code, and framework-specific configurations, making agents difficult to inspect, review, and govern. ADL enables portability across frameworks, transparent governance, and ecosystem interoperability while complementing existing standards like A2A protocols and Model Context Protocol.

5m read timeFrom allthingsopen.org
Post cover image
Table of contents
How ADL brings OpenAPI-style standardization to AI agent development.The problem: Why AI agent definitions are fragmentedWhat the Agent Definition Language provides for developersHow ADL relates to existing technologiesWhy ADL matters for open source AI developmentThe future of the Agent Definition Language standardHow to contribute to the Agent Definition LanguageMore from We Love Open SourceAbout the Author

Sort: