What is there to learn? You already know about Cyberhounds; which means you know that, without physical access to their brains, they are unhackable. They receive no external input besides IFF designators, usually laser based. Agents employing them tend to use laser designators to “Paint” targets, sending the Hounds after them. The Agents themselves wear radio badges that transmit a complex radio code that prevent them from being identified as targets.
You suppose you could hack them, if you were a robotics expert who knew more than the basics concerning this sort of IFF system and robot AI’s. As it is you don’t think you can fake the badges or the designators, not without days, probably weeks, of work. And hacking the Hounds directly is impossible, not unless you can plug into their brains.