Really high level stuff as far as I understood is done with proxychains, often with bulletproof hosts, hopped through several hostile/semi-hostile/uncooperative nations.
There are services in certain countries called “Bulletproof hosts” who’s only function is that they will not cooperate with law enforcement. They use techniques such as “fast flux DNS”, where all of the servers on their network are shuffling IP very frequently, and they intentionally keep no logs. This means even if law enforcement comes knocking with an IP, the host itself ought to not even be able to say *which* of their own servers did it.
A bulletproof host is paid with bitcoin through anonymous or stolen wallets, ideally.
The connections are chained through several of these, eg. turkey → russia → china → united emirates → istanbul, and the feasibility of pursuing the connection becomes extremely difficult/not feasible/not cost efficient. Get a warrant, go to istanbul, fight with the BPH, go to the united emirates, get a warrant, etc. This is a political and judicial nightmare.
The final hop is proper opsec. Never hack on your own network. Do not use any personal machines, buy a burner laptop, never connect it to *anything* you use. Do not log in to twitter, even if you wipe it after, do not connect to your wifi network, do not use a fancy custom browser setup, do not even turn it on at your house. This laptop’s sole purpose is to be as unrelated to you as possible to prevent data leak. A single correlated fingerprint burns the entire proxychain.
For remote CNC, techniques like anonymous twitter accounts can be used. Post a tweet that says “XYZ 123”, and your botnet that has been watching the twitter feed for instructions starts performing an attack. Same opsec applies. Variety can be used here, I have no experience in this, just enjoy reading up on it.