Re: shtml attachment files ?
Shawn Iverson via MailScanner <mailscanner-qhrM8SXbD5JTOyd/[email protected]> Mon, 26 Sep 2022 07:11:55 -0400
| Newsgroups | gmane.mail.virus.mailscanner |
|---|---|
| Message-ID | <[email protected]> |
No. The code that does this isn't even MailScanner itself. It is MIME::Parser when the Explode* methods are called. Changing this behavior would mean mostly likely rewriting that perl module. The problem you are running into is that the filename blocking is not MIME-aware and doesn't look at the email but rather the files that are exploded in the working directory. The distinction here is that these html files you are fighting are likely attachments and not the html mime part of the email. You have a few options: 1) Block any html file that doesn't conform to msg-<id>.html in your filename rules: allow msg-.*\.html$ HTML part of email message HTML part of email message deny \.html$ HTML not part of email message HTML not part of email message 2) Use SpamAssassin to examine the mimeheaders and score appropriately mimeheader Content-Disposition =~ /attachment; filename=\".*\.html\"/ describe HTML attachment, not safe! score 10.0 On 9/26/22 04:28, [email protected] wrote: > > > On Thu, 28 Jul 2022, Shawn Iverson via MailScanner wrote: > >> It is happening during the mime parsing operation when the mime is >> split into text and html parts. It assigns a name msg-<id>.html to >> the html mime part. > > > Is there any way to avoid that, can we assign a different name, other > than .html ? > > The number of .html attachments with links to a partial virus payload > and that are slipping through undetected is increasing and it's > worrying. I have a pretty decent proxy/filter but......... > > >> > > -- MailScanner mailing list [email protected] http://lists.mailscanner.info/mailman/listinfo/mailscanner