Clustering should be able to run without writing files

[hypercompetent]

We need to have a mode for clustering that doesn't write files to the users' machine. This is not normal behavior for most functions. The outputs aren't very large, so we should be able to contain these results in a list object.