The most successful attack would probably be whitelisting rather than blacklisting.  Whitelisting is synonymous with licensing or at least closely related.

To make maple syrup, one takes a bunch of tree sap which is nearly impalpable and slowly cook it down until is is of a suitable form to use on waffles.  This slow consolidation until a body consists of higher valued players who can be pressured is exactly what I see as the most viable form of setup for a successful attack, and matches fairly well a 20x increase then subsequent doublings in block size as is the current proposal.  It results in a corresponding slow 'boiling off' of smaller value support infrastructure.

The key balance here would be to work slowly.  Right now Bitcoin is not a significant threat to almost anyone so there is time.  It's more important to preserve Bitcoin's primacy in the space than it is to forcible destroy it because sudden moves will panic the herd into alternate solutions, and many of these could be much more challenging to attack both because of numbers and different designs.

I've not mentioned the likely effectiveness of attacks on fungibility for some time so I will do so again here.  Please note that an attack such as telling TigerDirect that they must follow a licensing regime would have broad ramifications across the space.  I personally am very much against loss of fungibility on philosophical and economic grounds, but I'm still going to personally de-value a 'tainted' coin.  That is, a spend which is not usable at TigerDirect because it's history cannot be traced through licensed transference.

Edit:  Note that I have no interest in spending my BTC at TigerDirect, but I still de-value BTC which cannot be spent there.  Why?  Because the guy I try to pay with them himself may wish to spend them at TigerDirect and they will be worth less to him.  Fairly simple economics.

Nobody has really made a credible argument to me that the U.S. govt could not or would not require Bitcoin users, and especially of those at the corporate retail level, to adhere to licensing (and thus implement the aformentioned fungibility attack.)  I personally see it as modestly trivial and some moves in that direction happening already.  Were I contracted by Uncle Sam to design an attack on Bitcoin, I would certainly employ the attack I mentioned, and I absolutely would not pull the trigger prematurely.  Today, especially prior to Gavin's hard fork, is very premature.