heres the problem, things like computerized visual recognition sucks. unless your talking about something like a QR code, it takes a lot of CPU power and a lot of time to code on top of being not very accurate. the way we interface with wow we dont really have to worry about warden anyway, i dont think its been a factor for the last few years in terms of wow botting. so its easyer just to read the values from the memory of the client and inject what we need to to make the bot operate properly rather then build a new bot that would be much worse then any of the AutoIt based bots people tried to make there a while ago.