> From: Mat Harris
> Sent: Sunday, February 08, 2004 7:21 AM
>
> Hi, I didn't get a response last time this was posted so here
> we go again :)
[...]
>
> It seems to be making the mails just slip through.
>
> Samples are found here:
>
> http://dev.genestate.com/debug/spam/
>
Although not a compelling spam score, it was enough to cross the
threshold, using some custom rules:
Content preview: Hey, man...havent had you take a drive by lately. Why
don't ya try sometime?:
http://sb3.chainedmicrophone.com/51074251/10488/1133/bigtime.html I
think if you'd just relent and see here, you'd be thrilled:
http://sb3.chainedmicrophone.com/51074251/10488/1133/bigtime.html [...]
Content analysis details: (5.4 points, 5.0 required)
pts rule name description
---- ---------------------- ------------------------------------------------
--
1.4 HTML_IMAGE_ONLY_06 BODY: HTML: images with 400-600 bytes of words
2.1 BAYES_90 BODY: Bayesian spam probability is 90 to 99%
[score: 0.9693]
0.1 HTML_MESSAGE BODY: HTML included in message
0.2 MY_TITLE BODY: Testing for HTML title in emails
0.9 MY_SHRT_IMG BODY: 1-3 letter gif or jpeg in url.
0.7 MSGID_FROM_MTA_HEADER Message-Id was added by a relay
The particular custom rules being hit, are from here:
http://www.merchantsoverseas.com/wwwroot/gorilla/oct03_rules.cf
It'd be nice if this sort of spam could be scored over 5.0 without Bayes,
but it'd probably require some additional custom rules unique to this type
of spam.
