I'm guessing that you may be referring to the includePackage command:
SparkR also allows easy use of existing R packages inside closures.
The includePackage command can be used to indicate packages that
should be loaded before every closure is executed on the cluster.
Source: R frontend for Spark
