Now perhaps a somewhat safer tool is to just use the cracking to determine the best targets to bribe or backdoor, but only allow the group with the cracking power to give the names of services to monitor at any cost.
I think what's most interesting along this lines is what happened during WWII when the allies cracked the enigma. Suddenly, they knew what the nazis were sending to each other. Bletchley Park had to keep most of the intelligence secret to itself, because the nazis could get wind of it and changes the procedures to encryption -- particularly if some top secret attack was somehow thwarted out of the blue.
That's why I said the part about "parallel construction". During WWII if the allies captured a spy or a high ranking officer, then they could maybe act on one piece of information -- giving the allies the necessary plausible deniability by blaming it on the captured nazi officer.
LEO and Prosecutors will use "parallel construction" to construct a narrative about how information was obtained in a legal way even though it was clearly obtained illegally.
Or you could choose to only act on 5% (e.g.) of the information gleaned -- and that which could clearly be shown to have a third party leak.
Or say if you were tapping the information of a mob boss, you could leak the information to a competitor and justice work it's way through the streets instead of the courts.