Using a combination of blacklist-style identification and Bayesian algorithms for identifying spammers from the contents of their tweets, this app from the folks at HashRocket adds another tool to our collective belts for fighting back against Twitter spam.
After authenticating your Twitter account, TweetBlocker generates a list of your followers with a numerical score as well as at-a-glance letter grades according to how spammy the service thinks they are, with A+ being the best and F the worst. From here you have the option to block or simply unfollow each account. Rolling over a user's icon pops up a list of their most recent tweets so you can quickly double check any accounts you might be on the fence about.
In our tests we did find some false positives in our list as well as some questionably low grades for relatively popular accounts (see above -- blip.tv passes with flying colors while Blip.fm seems dangerously close to repeating the fourth grade), but the recent tweets rollover function should help you make sure you're not inadvertently blocking or unfollowing someone having a momentary lapse of reason or temporary absence from Twitter.