I have a question

Modifying clients always gets you banned. There's no such thing as an "undetected exploit"; they can only hide the traces they leave which isn't foolproof. Sure they might not be actively detected, but they sure can with some RE, which is what Roblox does sometimes. The goal to safe exploiting is to manage risk, not use a so-called undetected exploit.

Manage the risk by using alts, connecting to a VPN (a bit overkill but why not if you pay for one like I do), VMs that bypass their VM detection, dual booting, or using a completely different physical machine. 

That's what exploit devs do to get past bans if they happen (well, at least me).