That is roughly how it works, self selection. The SF groups that, to put it charitably, operate under legally ambiguous circumstances are people who know each other really well, and they only invite in those who are a known quantity. There isn't any kind of laboratory-prisoner-execution/choose-the-red-pill sort of test, the evaluation occurs continuously in the daily performance of one's duties.
SEAL teams are, relatively speaking, high profile SF - and therefor much less likely to be tasked with strictly illegal and morally objectionable missions. I have a lot more trust in someone who has SEAL team affiliation towards the end of their military service. They've walked the hall and stopped opening doors pretty early on, so they're a well measured quantity.
Also, it is my experience that veterans with combat action ribbons are far less likely to mindlessly appeal to state authority or be swayed by utilitarian justifications. Of the dozen guys I've kept in touch with, most are libertarians and a few are anarchists. I guess seeing the ugliest side of the state will flip a switch :)
I know that some would say no. I don't know how to tell ahead of time. It's a hard, general problem that's all probabilities and personality. Yet, with them, it's even harder.
That is how you end up with a spectrum of SF groups, where some are tasked with difficult but clearly whitehat missions (the bulk of well known SF), illegal but debatably moral grayhat (some known SF but mostly units operating under cover designations), and finally the blackhats that rise to Hollywood level evil (basically Apocalypse Now with no paperwork). So for a late career SEAL the probabilities are pretty good for someone to land between lawful-good and neutral-good. Those are infinitely better odds than what you'd get with some random businessman.
I'd probably share your concern if the military was full of utilitarian statists, but strangely enough it is the direct opposite... well, except for flag officers - they can get a little kooky.
You're implying the Old Boy network effect doesn't exist. Whereas you discuss it in other posts. This is a concern to me even if I accept the arguments about what their views are likely to be on average walking out of the SF community. When the effect applies, they're often working for each other or previous groups as much as the current one.
Kind of, but not really. One is monolithic and the other is stratified in such a way that loyalty does not transfer to the superstructure. Consider the relationship between the US Navy and Red Cell - a perfect, and largely public, example of the compartmentalization and transferability of loyalty.
> ...or previous groups as much as the current one.
And since the order of potential group membership and level of potential shadiness is so well defined - that is a pro rather than con when the last group to which one was a member can be determined.
Regarding calling shots, yeah who knows. Just reinforces worries in this space though. Uncertainty that is.