inputAction
The action to take when harmful content is detected in the input. Supported values include:
BLOCK
– Block the content and replace it with blocked messaging.NONE
– Take no action but return detection information in the trace response.