The real way to fix security issues isn't to find new exploits and expose them. It's to architect new ways to prevent whole families of exploits from being possible. I've only really seen Project Zero do one sort of recommendation in this way. Outside of that, I don't feel that them finding 0days and releasing them is actually a significant improvement on security because they aren't even close to plugging all of the holes (or even enough to really make a difference).
In terms of raw "national security" net gain, it's likely it would be more useful if kept secret and used offensively than if it was revealed and patched. Vulnerabilities are dime a dozen. Patching one zero day says nothing of the 10 more that are floating around at any given time.
Back around the year 2000 Microsoft was being beaten up for all the vulnerabilities in their software. So in 2002 Bill Gates announced "Trustworthy Computing".[1][2]
In 2014 Microsoft finally threw in the last towel, folding the group they formed into other units. They gave up. Microsoft lost not because they were incompetent, but because the problem is too big to attack in a conventional manner.I don't know what the answer is, but we need to approach things very differently. To quote Dr. Peter Venkman: "the usual stuff isn't working".
[1] http://www.foxnews.com/story/2002/01/16/bill-gates-announces... [2] https://en.wikipedia.org/wiki/Trustworthy_computing