Looks connected to me

Being on the LAN is just a quick way to say low latency + high capacity, some threshold for one/both of those could be where the desync happens.
It's like the 30 FPS limit for research success, it probably needs a packet rate limit (like "trickle" in Linux, probably most firewalls can do it).
It sounds like the target CMDR is hosting the scout control, then jumps/exits without passing over the control (exit may work better than jump there).
You're counting on the transfer to fail - if it works, the scout script runs and they either fight if they can see the other CMDR heat per distance or leave it not.
A failed transfer leaves the other client knowing where the scouts are but not controlling them, though I think that limbo state can resolve if more CMDRs leave/join.