Since proof-of-concept code for the JPEG flaw in Windows has been posted online, we can surely expect at least one mass mailing exploit soon. The form will likely take the form of either:
1) A JPEG file embedded in an email message with the exploit code embedded in the embedded image. Theoretically, the exploit pattern should already be known, no matter what the encoding is, so anti-virus companies should theoretically be able to detect this already, if this method is used.
2) Because of the above, the more likely method seems to be the embedding of a URL in the message that either refers to the actual JPEG itself or refers to a webpage that loads the infected JPEG. It seems then that the only tool that could detect worms of this sort would be SURBL.
And so on to my question: if I (or anyone else for that matter) submit a domain name that hosts an infected JPEG file, how quickly will the SURBL databases be updated to reflect this infection?
Also, what if the exploit is multi-stage, and tries to infect actual http servers with infected JPEGs, and thousands of websites become infected...? Would it then be necessary to create a separate SURBL list for these infected domains, or could they be listed in, say, the phishing list?
Thanks, Matthew Wilson