Exactly. Tools to be used in a proxy-war with little thought on long-term consequences.

Eh, not really. There is a whole group within the Pentagon that evaluates long term strategy, J5. The problem is, it's impossible to tell what will happen in the years after a particular conflict subsides on either side.

I could've told you whether a coup was likely to increase or worsen stability: always default on the latter if imperialism is the motive. Likewise, financing dictators' wars or religious nutjob's training in beating superpowers. Either this group you mentioned is incompetent or nobody listened to them. I'm going with the latter since many of the documents warn of high risk of blowback.

Except you would probably be wrong. I know the crowd here likes to think they can solve all problems better than the status quo but the reality is that for these kinds of things data doesn't help much. See: COIN