Unity Gateway Skills API and developer tools available in Beta

The Unity Gateway Skills REST API and its corresponding developer tooling are now available in Beta. This enables engineering teams to programmatically automate the complete lifecycle of skills for AI agents.

Use cases

Imagine you have a set of instructions and routines that your AI agents need to reference, and you want to automate publishing and updating them through your CI/CD pipelines.

• Create, list, update, read, and delete skills programmatically.
• Upload instruction files in Markdown format to Unity Catalog storage.
• Validate and finalize skill bundle registrations.

How to use

Creating a skill requires creating the resource, uploading its bundle, and finalizing the operation. Here is an example of creating a skill using the Python SDK:

from databricks.sdk import WorkspaceClient
from databricks.sdk.service import catalog as c

w = WorkspaceClient()

skill = w.ai_gateway.create_skill(
    parent="schemas/main.default",
    skill_id="my_skill",
    skill=c.Skill(comment="Team-shared math helper skill"),
)

What happens in this code:

1. Instantiates an authenticated Databricks workspace client.
2. Calls the skill creation method passing the parent schema and desired skill identifier.
3. Sets a descriptive comment for the new shared skill.

Limitations and requirements

• This feature is in Beta and can be managed by account admins from the account console Previews page.
• Developer tools require specific minimum versions (for example, Databricks CLI v1.20.0 and Python SDK databricks-sdk 0.142.0).
• Skills do not have Terraform or Declarative Automation Bundles resources.