For agents, the value is clearer still: structured JSON output, reusable commands and built-in skills that let models interact with Workspace data and actions without a custom integration layer.

My LLMs pair incredibly well with these tools ...