Followup on guardrail config

Generally seems good.

We could consider changing extra_body={"guardrails": True} to be more specific: extra_body={"moderation": True}. Or consider nesting to allow for additional guardrails to be added in the future: extra_body={"guardrails": {"moderation": True}}.

Should the guardrails default to enabled if an endpoint has been configured?

Originally posted by @eoinfennessy in https://github.com/ogx-ai/ogx/pull/5291#pullrequestreview-4227455670