Attribute expansion allows you to enrich agents in the AI Agent Registry with additional metadata beyond the default and vendor-provided Attributes. This enables you to capture business, operational, and governance-related information that is specific to your organization.
By extending the Attribute model, you transform the registry from a basic inventory into a structured dataset that supports advanced use cases such as dynamic grouping and future Policy enforcement.
All additional Attributes apply to every agent in the registry, regardless of whether the agent was discovered through integration or registered manually.
Create Additional Attributes
To create an Attribute:
- From the Environment navigation bar, hover over AI Agent Registry and click the settings icon.
- Click Attributes.
- Click New Attribute.
- Enter the Attribute details, including name, type, and any constraints such as length or allowed values.
- Configure whether the Attribute is available for Policy usage or access requests.
- Click Save.
The Platform supports multiple data types, including string, numeric, and Boolean formats. This ensures that you can model a wide range of metadata.
Once created, the Attribute is added to the registry schema and becomes available across all agents.
Managing Additional Attributes
To use additional Attributes:
- Click AI Agent Registry in the Environment navigation bar.
- Hover over the relevant agent and click the Edit icon.
- View and update the additional Attributes in the side panel.
- Click Save.
Attribute Behavior
Additional Attributes are fully editable and can be removed if no longer needed. They are populated manually and are designed to support flexible data modeling.
When a new Attribute is created:
- It is automatically applied to all agents in the Workspace.
- It becomes available for use in dynamic grouping and future Policy definitions. Ensure that the Attribute is available for Policy usage or access requests.
- It is included in the registry schema and mapping model.
Example Use Case
You can use additional Attributes to classify agents based on organizational needs. For example, you might create a Boolean Attribute such as PII Access to indicate whether an agent interacts with sensitive data, or an Active Attribute to track operational status.
These Attributes can later be used to group agents dynamically, such as identifying all agents that access sensitive data.