To understand and make use of nFoldMonOnly, you should already have, and thoroughly understand, nFoldMan.
nFoldMonOnly is merely a version of nFoldMan without the Mac user interface, and configured to run as a unix-level command, not an application. As such, it can be run (only) from the Terminal program, or as a unix StartupItems script. Since it is not an application, it does not take up space in the Dock, and cannot be accidentally Quit by a novice user interacting with the visible Mac applications. (It can still be stopped from Activity Monitor, or from a Terminal command.)
Although nFoldMonOnly has no Mac user interface, it will respond to network connections from instances of nFoldMan, exactly like nFoldMan does (in fact, it is the same code as nFoldMan). A simple control file defines the folding processes nFoldMonOnly manages, and all interaction (status display and control signals) are via a copy of nFoldMan running somewhere - on the same machine or elsewhere on your network.
In my typical application, then, I add a StartupItems script to a workhorse machine. This StartupItem starts nFoldMonOnly and leaves it running permanently. nFoldMonOnly is configured to start the folding client when the machine is idle for more than some period (10 minutes in my setup), and to stop the folding client as soon as the machine becomes active again. The result is that the users of these lower-powered workhorse machines never compete with the folding process, and have no opportunity to interfere with the monitoring process. A copy of nFoldMan running on another machine is used to monitor the status of folding across our home network.