Today, a security vulnerability in TweetDeck allowed malicious code to be run remotely. It seems the primary symptom is your account retweeting the malicious code. To fix it, sign out and back in now.
The code runs automatically as soon as it appears in your stream, so you don't have to interact with it to be affected. Twitter has said that the vulnerability is fixed and you only need to sign out of your TweetDeck account and sign back in. Twitter has also taken down TweetDeck services, so logging back in might not work just yet.