AWS Backs Agentic Resource Discovery, an Open Specification for AI Agent Discovery

What AWS announced

On August 24, 2026, AWS published details of Agentic Resource Discovery (ARD), an open specification designed to make AI-agent resources discoverable across environments. AWS positions ARD as a shared catalog and discovery format that can describe agents, tools, skills, MCP servers and APIs without forcing every platform to build bespoke registry-to-registry connectors.

How ARD fits with AWS Agent Registry

AWS Agent Registry provides a centralized catalog for resources inside an AWS environment, including curation workflows, hybrid semantic and keyword search, MCP-native access and IAM or JWT-based authorization. ARD extends the idea beyond a single environment by defining a common way for registries to describe and expose resources. AWS compares the federation goal to the way DNS lets independently operated networks participate in a common discovery system.

Why interoperability matters

As organizations deploy agents across clouds, on-premises systems, SaaS applications and developer tools, discovery becomes a governance problem as much as a convenience problem. Teams need to know what an agent or tool does, who owns it, whether it is approved, and how clients can locate it. A portable specification could reduce duplicated integration work and make it easier for policy-controlled catalogs to interoperate.

Open-standard status

AWS states that ARD is an open standard rather than an AWS-only product and that the specification is available under the Apache License 2.0 with reference implementations. AWS says it contributed feedback during the specification's development. That distinction matters: the long-term value of ARD depends on adoption beyond a single vendor and on whether independent registries implement compatible discovery behavior.

What to watch next

The key signals will be implementations outside AWS, interoperability testing among registries, security and trust metadata, adoption by MCP-compatible clients, and real enterprise use of federated discovery. If the ecosystem converges, ARD could become an important layer for locating governed agent resources across otherwise fragmented AI environments.