aiAgentGuardrailUpdate

Update an existing Guardrail in the AI Agent Registry.

Description: This mutation updates an existing guardrail associated with the authenticated user's account.

You can update an existing Guardrail and associate it with multiple new agent versions. This association is an incremental operation, meaning it adds new agent versions and skip any that are already associated.

When a Guardrail is created, it's initially associated with an agent version, for example, with GUID <xxxxxxx-1234-xxxx-xxxx-xxxxxxxxxxxx>. Upon updating the Guardrail, this existing agent version GUID is retained, and the Guardrail is also associated with a new agent version, such as <xxxxxxx-4567-xxxx-xxxx-xxxxxxxxxxxx>.

The response demonstrates this incremental association, showing both the previously existing and the newly associated agent versions.

Input Parameters (AiAgentGuardrailUpdateInput):

• id: ID! - Required. ID of the guardrail to be updated.
• name: String - Optional. Name of the guardrail.
• description: String - Optional. Description of the guardrail.
• configurationJson: String - Optional. Updated JSON configuration for the guardrail detailing its policies and operational parameters.
• status: AiAgentGuardrailStatus - Optional. The operational status of the guardrail (e.g., ENABLED, DISABLED).
• agentVersionIds: [String!] - Optional. Agent version IDs associated with this guardrail.
• externalId: ID - Optional. External ID for the guardrail.
• tags: [AiAgentTagInput!] - Optional. Tags for categorization.

Response:

• Returns the updated AiAgentGuardrail object.

Example Request:

mutation {
  aiAgentGuardrailUpdate(input: {
    id: "guardrail-123",
    name: "Updated Data Privacy Guardrail",
    description: "Prevents access to sensitive and confidential data.",
    configurationJson: "{ \"priority\": \"high\", \"category\": \"data-processing\" }",
    status: ENABLED,
    agentVersionIds: ["v1", "v2", "v3"],
    externalId: "guardrail123",
    tags: [
      {
          key: "security",
          value: "critical"
      }
    ]
  }) {
    id
    externalId
    name
    description
    configurationJson
    status
    tags {
      key
      value
    }
    agentVersions {
      id
      name
      version
    }
    auditData {
      createdByUserId
      createdTime
      modifiedByUserId
      modifiedTime
      updatedByOrigin
      updatedAtProviderTime
    }
  }
}

aiAgentGuardrailUpdate(
  input: AiAgentGuardrailUpdateInput!
): AiAgentGuardrail

Arguments

aiAgentGuardrailUpdate.input ● AiAgentGuardrailUpdateInput! non-null input agent-control-tower

Type

AiAgentGuardrail object agent-control-tower

Represents a guardrail in the AI Agent Registry system that defines policies and operational parameters for AI agents.

Guardrails are security and compliance mechanisms that control AI agent behavior by enforcing rules, filtering content, and ensuring adherence to organizational policies.