How to identify the element(s) of an email that triggers the scam detector
I know how scam detection works. Sometimes an email from a legitimate source is flagged as a scam. That's ok, but sometimes I need to identify exactly the point(s) where the scam detector found the problem, in order to suggest the sender to change. Often it is a simple typo in the footer that makes the link different from the text, and it is good if the sender changes its footer.
The point is that, to do this, I need to open the email source and visually replicate the job done by the scam detector. Is there a way to spot the problem with the help of the scam detector itself? An Add-on maybe?
Additional System Details
- User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:66.0) Gecko/20100101 Firefox/66.0
There is a bit of information here. If you want to dig deeper, this recent bug may offer a starting point. There are a few discussions from ca. 2005, but I don't know how many of them are still relevant.
Thanks for your reply. I have already seen that page and other describing how scam detection works. This is not the point. What I want is an help to spot the single wrong link is 300K+ email. Doing this by hand is almost impossible.
What I would like to have is an hint from the scam detector, on where it has found the problem.
Is this possible? Having Thunderbird underline it with a wiggle would be the best, but even looking at log files or other internal information can be useful.
As you have found, there appears to be no public, documented description of the scam algorithm, let alone a tool to apply it to bulk quantities of mail in the way you describe. Your best option may be to contact the participants in the Bugzilla reports.