[Rspamd-Users] Block emails with Chinese and Cyrillic characters in the subject

[Andreas Günther]

Am Mittwoch, 6. August 2025, 16:08:19 Mitteleuropäische Sommerzeit schrieb Steve 
Witten:
> Here are the relevant regexps in
> $LOCAL_CONFDIR/local.d/maps.d/blacklist/content.map:
> 
> # Non-Western charsets
> #
> /(=\?|charset\=)big5/i                  # Big5
> /(=\?|charset\=)euc-kr/i                # Extended Unix Code -- Korean
> 
> /(=\?|charset\=)windows-1251/i          # Windows 1251
> /(=\?|charset\=)koi8-r/i                # Cyrillic
> /(=\?|charset\=)gb2312/i                # Chinese
> /(=\?|charset\=)ks_c_5601-1987/i        # Hangul (Korean)


These are outdated rules. They're barely effective anymore because they recognize 
outdated language-specific character sets like windows-1251, gb2312, or big5, while 
modern spam emails almost always use universal UTF-8.


That's why my solution is this:


/etc/rspamd/local.d/multimap.conf
MIXED_NONLATIN_SUBJECT { 
   type = "header"; 
   header = "Subject"; 
   map = "${LOCAL_CONFDIR}/local.d/maps.d/mixed_nonlatin.map"; 
   symbol = "MIXED_NONLATIN_SUBJECT"; 
   description = "Cyrillic, Chinese, Arabic or other non-Latin subject"; 
   regexp = true; 
   score = 6.0; 
   action = "reject"; 
}

and in 
/etc/rspamd/local.d/maps.d/mixed_nonlatin.map

/\p{Cyrillic}/u 
/\p{Han}/u 
/\p{Arabic}/u 
/\p{Hangul}/u 
/\p{Hiragana}/u 
/\p{Katakana}/u


Best regards


Andreas