Hi,

A new phishing detection algo written by me as a part of the Google Summer of
Code program is available in the cvs version of clamav.

Documentation on daily.pdb/daily.wdb, is available in docs/phishsigs_howto.pdf
Algorithm docs are at: http://wiki.clamav.net/index.php/phishing_design.

The phishing module processes pairs of realURL/displayedURL.
I) If they match against a whitelist (daily.wdb) the urls are clean.
Otherwise further proccessing is done. There are two modes of operation:
       1) ----
               The phishing detection module matches
realURL/displayedURL pairs against
               daily.pdb. The url is processed further only if it matches.
       2) alldomains
               daily.pdb is ignored and all urls are processed
II) It is determined if the realURL/displayedURL is a phishing attempt.
III)  The returned possible "virus" names begin with Phishing.Email,
and are: HexURL, Cloaked.NumericIP, Cloaked.Null,SSL-Spoof,
Cloaked.Username.

To enable build with ./configure --enable-experimental
and use the --phish-scan-alldomains param to clamscan or  the
PhishingScanAllDomains config option in clamd.conf

To disable:
use --no-phishing-scan-urls, and the corresponding PhishingScanURLs
option in clamd.conf

Looking forward for you ideas, comments, results (and bug reports).

Best regards,
Edwin
_______________________________________________
http://lurker.clamav.net/list/clamav-users.html

Reply via email to