derhansen/logwatch-modsec2 on GitHub has a great example of a Logwatch filter for mod_security. The email output it generates looks like:

--------------------- ModSecurity2 (mod_security2) Begin ------------------------

ATTACKS BLOCKED ON VHOSTS:

subdomain.domain.tld - 2 time(s)
[ip: xxx.xxx.xxx.xxx] [id: 981231 ] [msg: SQL Comment Sequence Detected.]  - 1 time(s)
[ip: xxx.xxx.xxx.xxx] [id: 981231 ] [msg: SQL Comment Sequence Detected.]  - 1 time(s)

www.site.tld - 1 time(s)
[ip: xxx.xxx.xxx.xxx] [id: 990012 ] [msg: Rogue web site crawler]  - 1 time(s)
[ip: xxx.xxx.xxx.xx] [id: 981318 ] [msg: SQL Injection Attack: Common Injection Testing Detected]  - 5 time(s)
[ip: xxx.xxx.xxx.xx] [id: 950901 ] [msg: SQL Injection Attack: SQL Tautology Detected.]  - 2 time(s)

www.anothersite.tld - 1 time(s)
[ip: xxx.xxx.xxx.xxx] [id: 958291 ] [msg: Range: field exists and begins with 0.]  - 1 time(s)

TOP 10 BLOCKED IPS:
xxx.xxx.xxx.xxx - 2 time(s)
xx.xxx.xxx.xxx - 1 time(s)
xxx.xxx.xx.xx - 1 time(s)
xxx.xxx.xxx.xx - 1 time(s)
xxx.xxx.xxx.xxx - 1 time(s)

---------------------- ModSecurity2 (mod_security2) End -------------------------

If you’re just looking for an example of parsing the log file, the perl script that does that is here.