Yesterday, Websense Labs issued an alert regarding a compromised CNET blog, namely the Clientside developer blog
"Websense Security Labs ThreatSeeker Network has discovered that a CNET Networks site has been compromised. The main page of the CNET Clientside Developer Blog contains malicious JavaScript code that de-obfuscates into an iframe that loads its primary malicious payload from a different host. The malicious code is observed to exploit a known integer overflow vulnerability in Adobe Flash (CVE-2007-0071). At the time of this alert, the site is still hosting the malicious code. Visitors who are not patched against this vulnerability will be infected without any user interaction."
Interestingly, the second javascript obfuscation that they analyzed in the time of detection is different than the one I managed to obtain from a copy of the blog on the 2nd of August. And while it remains unknown for how long has the blog beed embedded with the javascript with the, this malware attack, and the rotating javascripts indicate a compromise compared to the massive SQL injections we're seeing on daily basis. The embedded javascript code appears to have been removed. Deobfuscating the obfuscated javascript code, attempts to access the live exploit URL from a .info domain that is now down. Historically, the same domain has been used in blackhat search engine optimization campaigns - yet another example of underground multitasking, namely, abusing a single domain for several different fraudulent purposes.
Slowly, but inevitably, the "do no visit unknown and potentially harmful sites" security tip is starting to lose its charm.