What changed
AWS has added custom metadata schemas to AWS Agent Registry, allowing organizations to define and enforce standardized attributes across records. This feature enables administrators to create schemas that specify required and optional fields, along with their data types and allowed values. These schemas are applied to the registry, and all records must adhere to them. Publishers must supply values for the fields specified in the schema, and AWS Agent Registry validates these values against the schema set by the admin.
What this means in practice
Organizations can now use standardized vocabulary to describe their resources, ensuring consistency across records. The registry enforces the schema on every record, so fields like environment, businessUnit, and securityScanStatus have the same name and allowed values regardless of which team publishes the record. Custom metadata values are indexed for search, allowing consumers and agents to filter records using natural-language queries combined with specific metadata filters. Only approved records are returned, ensuring that records matching filters but not yet approved are excluded from results.
Limitations and availability
The feature is available in AWS Agent Registry and requires an AWS account in a supported AWS Region. Users need AWS credentials with Boto3 and IAM permissions for AWS Agent Registry. Python 3.11 or later and boto3 1.43.113 or later are also required. The sample implementation is available in the GitHub repository. The feature is currently in preview and requires manual approval for records, as the registry does not auto-approve any submissions.
