On Thu, 2016-11-17 at 20:49 -0500, Alex wrote: > Hi all, > > We have a lot of users who use email to share photos. Empty body, 2M > JPG attachment, nothing in the subject. > > How do you train bayes for these? > > Is 3.0 a suitable score for BODY_EMPTY? That seems quite high. I also > don't recall seeing it that high in the past, or it effectively being > enough to push the email into spam. > Sounds like a case for a local rule to me: one subrule to check that the body is *only* a JPG attachment and another making sure that the subject is blank.
> It's also a great way for a spammer to then sending image spam. I'd > appreciate any ideas on how best to manage this. > Indeed, though I haven't seen that type of spam for a long time and IIRC they tended to carry obfuscatory text as well as the image payload. Martin
