> I don't believe that Snowden-like reveals put him in any more specific danger
That's true only in the same way that sending half-rations to your soldiers on the front line instead of full rations doesn't put them in "more specific" danger.
Or in a civilian context, refusing to wear seat belts doesn't put people in any "more specific" danger, nor does smoking cigarettes. Yet we understand that statistically speaking across the entire population in question, that it is possible to reduce the risk of harm by taking specific actions that have general impact.
In the case of the military, it's always a double-ended question. There is the risk added or reduced for your force in particular, but also the risk added or reduced for the adversary forces.
E.g. you put "insurgent" in scare-quotes when mentioning that they actually pose a danger to your family in the military, but without seeming to realize that many of the NSA programs Snowden leaked are directly aimed at those very same insurgents putting your family at risk. Even if we take it as true that the U.S. managed to unilaterally create insurgencies (which is by no means the case), that is not the fault of U.S. servicemembers like your family any more so than any other non-U.S. politician.
To loop back to my point about managing general risk, the things that Snowden leaks had disastrous real effects, including second- and third- order effects. Beyond improving the security posture of U.S. insurgent adversaries and terror networks (terror networks that strike against U.S. allies in addition to the U.S. itself, I might add), Snowden's leaks also had the effect of weakening the national intelligence agencies of important U.S. allies, and reducing Western nations to infighting even as they needed to be able to present a united front due to geopolitical changes. It's hard to make statements about past "what-ifs" (though that has never stopped Snowden's supporters from chicken-littling about future what-ifs...), but certainly the joint U.S./E.U. reaction to Russia's invasion of Crimea would be more meaningful without Snowden's actions a year earlier.
But just because these risks (for the U.S. and for the possible adversaries or the U.S.) can't be quantified or laid out in short and simple flowcharts doesn't mean that those risks don't exist.