I suspect that the auto-start logic is making things needlessly complicated. When a serial port is setup as being a "JeeSketch" (I know, the names get a bit tedious), then JeeMon waits for an incoming "blah]" line and tries to start the "blah/host.tcl" script if present. My thinking was that I could plug in whatever JeeNode or Arduino I wanted, and JeeMon would launch the matching script to deal with it.
Then came the wireless part, whereby you have to config that say "node 12 is running a 'blah'" sketch. Then, the moment a packet comes in from node 12, JeeMon again launches the matching script. So turning on a node is like plugging in a USB device.
All nice and well, but for now I think it just makes the whole startup mechanism a bit too magical (and hard to get right, initially).
Bear with me. Let's see if this would be a practical way to stabilize and document things:
did you see the examples/ dir in JeeMon? - three examples, simple, but perhaps not that useful...
howto #1 could be: hook up JN, upload sketch to it, launch JeeMon, see serial output
howto #2: upload modified sketch to turn it into remote JN, hook up 2nd JN or JL running "central" sketch, launch JM, see incoming packets
howto #3: add decoders, to turn incoming packets into msgs such as "VOLT xys", etc?
howto #4: ... ?