I don't know whether this command is actually implemented or not. However, I can say that it is completely worthless. The extension adds nothing the protocol can't handle today already. The client can do this already by sending a D-type (or E-type) message; "DCMD ...". The other clients in the hub should not care about the type, as required by the spec. The hub however, should only forward D-type CMDs if the users are trusted, otherwise anyone can send them (as we saw recently). So basically, just use a D-type CMD.This is a protocol extension idea to allow a client or bot to send data to other users seeming to come from the hub:
client sends to the hub: HRLY <targetsid> <cmd> <data>
hub sends to targetsid: <cmd> <data>
This can be used for example to allow an external bot to send usercommands:
bot sends: HRLY ABCD ICMD <usercommand>
hub sends to ABCD: ICMD <usercommand>
I can't really think of any other command that would a) require that the hub is seen as the sender and b) can't be done through the use of a D/E type message.