We've all read about how Edward Snowden got access to thousands of classified documents and shared them with reporters. But how did he do it? He got access to tons of documents that he was not cleared for, i.e., that he should not have been able to access. What do we know about how he was able to access them?
Motivation. I think this case study might have fascinating insights for security against the insider threat. It might help us get a better understanding of what the threats are and how to defend against them. We've all heard the Santayana quote: "Those who cannot remember the past are condemned to repeat it". And, of course, if an insider can get this kind of access inside the NSA -- which I'd hope is one of the most locked-down environments around -- that should really teach us something about the insider threat in other environments. So, I think it'd be helpful for us to understand what technical methods Snowden used to access documents he wasn't authorized to access.
Clarification. There's lots and lots of speculation out there. I want to clarify that I'm looking for what we know about how Snowden did it, rather than your personal guesses/speculation. I would ask that you avoid speculation, and instead stick to facts, citable references, reports from those in a position to know, etc. If you absolutely cannot restrain yourself and must include some speculation in your answer, please make sure that you clearly distinguish between speculation (hypotheses, inferences, etc.) vs what's known (e.g., information that's been confirmed by someone in a position to know).
For instance, it's been widely reported that Snowden was a system administrator, which invites the speculation that maybe he used his administrator privileges to access documents that were being viewed by other users on machines he administered. But do we know that, or is that just speculation? It's also been widely reported that Snowden had some training in security and exploitation, which invites the speculation that maybe he used some sort of technical attacks to defeat certain security mechanisms. But do we know that, or is that just a natural hypothesis? What do we know?