by EKjellquist » Tue Mar 21, 2017 1:28 pm
Been looking at DKIM also; we have one particularly annoying source of spam that shifts enough to avoid RBL hits or direct blacklisting (because the domains end in *.us) and foils the bayesian filter to a degree with images and body text semi-coherent word salad. I was also thinking about using the SPAM-SPF-NONE-NEUTRAL variable to content-filter anything w/o SPF records into account quarantine folders automatically...