Back during the 'cold war', '62 or so, I produced data that went into an occasional document that was "classified" because it was an embarrassment to our government.

In this case the 'nuclear secret' might just be that something did not work.

Yea, I think a lot of what gets classified (then and now) is stuff that would embarrass or implicate high ranking officials if released.

<link>

Jane,

One of the last things you want your adversary to know is WHAT DOES NOT WORK. For instance, plutonium devices rely on very large numbers of things that MUST go right. If I am not mistaken, N. Korea had a plutonium dud test of its own.  If it was a Uranium device that did not work then they are really bad at it. Or, perhaps, spoofed.