Search results

  1. L

    Testing scoring with ham/spam

    Just out of curiosity, how do you test it via PMG directly? I know how to do it via spamassassin, but if I can test it via PMG (somehow), then obviously I can try that - my result without KAM_BITCOIN is simply an email from earlier that went into quarantine Best Regards, Lucas R
  2. L

    Testing scoring with ham/spam

    # less /usr/share/spamassassin-extra/KAM.cf | grep __KAM_INFOUSMEBIZ1 rawbody __KAM_INFOUSMEBIZ1 /http:\/\/(?:www.)?.{4,30}\.(info|us|me|me\.uk|biz)(?![-\.])(\b|\/)/i meta KAM_INFOUSMEBIZ (__KAM_INFOUSMEBIZ1 + __KAM_INFOUSMEBIZ2 + __KAM_INFOUSMEBIZ3 >= 1) It indeed won't...
  3. L

    Testing scoring with ham/spam

    hmm, from my spamassassin command on the exact file (wget it onto my PMG server): root@pmg1:~# cat spam.txt | spamassassin -D .... snipped ... Received: from localhost by pmg1.cluster.eu with SpamAssassin (version 3.4.6); Fri, 21 May 2021 17:20:16 +0200 From: DailyCrypto News...
  4. L

    Testing scoring with ham/spam

    It's attached! Sorry for the late reply - I missed the notification - Lucas
  5. L

    Testing scoring with ham/spam

    I'm guilty here for TXREP - I implemented it in favor of AWL, in my specific case it seems to get more accurate in fixing judging mistakes quicker (both for ham and spam) than AWL did - I know it requires actual training, and I'm aware it modifies what PMG ships with For the sake of it...
  6. L

    Testing scoring with ham/spam

    Hi @hata_ph Thanks for your reply! I'm aware of the condition itself, but the same email going through PMG and being manually tested with spamassassin gives two different results, one (spamassassin direct) triggers the rule, the other one does not - I would expect the condition to be equal...
  7. L

    Testing scoring with ham/spam

    Hi! Is there any way we can run a test to check the score a given email would get (ham or spam). We can call spamassassin directly like: cat /var/spool/pmg/cluster/1/spam/1A/12075460A6EABB8A01A | spamassassin -D pyzor | less This will tell us the assigned score (and the info) - but can we do...
  8. L

    Allow fetching quarantine link via API call

    Hi Dominic, I'll create a PMG Bugzilla! Fun fact, the email notification from this forum, somehow got marked as spam by PMG :oops: Main trigger being KAM_MAILBOX2: KAM_MAILBOX2 6.25 Mailbox Quota Phishing Scams Have an awesome day!
  9. L

    Allow fetching quarantine link via API call

    Hi guys, Currently, a user can request the quarantine link via the https://pmg.proxmox.com/pmg-docs/api-viewer/index.html#/quarantine/sendlink call - which sends them an email with the link to log into quarantine. Would it be possible to add an authenticated version of this (so e.g. for...