The argument setter plugin available in the Hub has a horrible reference doc that just says:
Performs an HTTP request some endpoint to get a driver specification. Based on the spec, it will make another call to a nebula endpoint to get data about the dataset, which it will use to set 'input.path', 'input.properties', 'directives', and 'output.schema' arguments that can be used later on in the pipeline through macros.
This is not at all how the plugin works. The github repo (https://github.com/data-integrations/argument-setter) has a README that has a much better description.
Fixed the documentation and widgets for the Argument Setter plugin in Hub