The first few times it's going to be expensive, but once everyone level sets with intense scans of their codebases, "every single release" is actually not that big a deal, since you are not likely to be completely rebuilding your codebase every release

You still have to account for the non-deterministic behavior of an LLM, when do you know you have exhausted its possible outcomes for any given piece of code?

I'm not sure. An innocuous one line change like "bump version" possibly adds a million new lines of code.