I am having this oft-reported problem of large amounts of spam suddenly coming through. There is a particular new style of spam which it seems is causing the problem - the names and links change but theory are all formatted the same way. According to the Log, SS is auto-adding the senders to the Whitelist! And I can’t think why it would do this. (I have recently migrated my whole system and files to a new disk, but SS is apparently still working in Apple Mail.)
Typical log entry:
Trained: Good (Auto)
Subject: Hi from Eileen
Identifier: v5ZFupdWe9SCux8oDyIF7g==
Actions: added rule <From (address) Is Equal to "eileen_garcia@celadjuvenil.org"> to SpamSieve whitelist, added rule <From (name) Is Equal to “Eileen Garcia”> to SpamSieve whitelist
Same problem, checked the FAQ …
SpamSieve is auto-training incorrectly. I’ve checked my setup, and I correct SpamSieve’s bad predictions promptly, but it’s still putting spam on the whitelist.
It is normal for SpamSieve to put the addresses on the whitelist. If the prediction turns out to be bad, SpamSieve will uncheck whitelist rule. Are you seeing different behavior?