That's no different to requiring an arcane incantation of shell scripts that call maven, git, curl, docker, , etc. The advantage of having it built into IDEs or CI servers (and by built in I mean TeamCity/Jenkins providing it, not a self inflicted third party plugin) is that you can build around these steps. Imagine if all CI servers supported merge previews or mergebots - that workflow gets unlocked for everyone!
at least with the shell scripts I can quickly see what it does. I much prefer the unix philosophy of small bits of code that do one thing, and do it well. Such things can be combined to make bigger things. Want to make a plugin out of those individual scripts? Knock yourself out! Meanwhile, a decade from now, some engineer can actually reconstruct if needed. Much better than a 'binary blob' plugin that gets abandoned.
