You can test procmail in a number of ways. The way I found easiest was just to simply send an email I was testing against to stdin for procmail

$procmail < email

The danger here is that you are running and testing your scripts on a live system, and hence if something is wrong with your recipes, bye-bye messages.

If you want to be really safe though, you add a safety catch recipe, as your first rule:

:0c
/some/backup/dir

Not actually my own tip. Taken from David W. Tamkin's Post on the Procmail Mailing List.

The only problem is… it doesn't work. At least not for me.

I finally got the code working by messing around a bit with it.

DROPPRIVS=yes
:0fw
| /usr/bin/spamassassin
:0
* ^X-Spam-Status: Yes
{
  # where i want the message to go
  foldername=$HOME/Maildir/.zzSpam

  #create a carbon copy of the message and store it
  #as a non-delivery action, this allow the script to
  #continue processing see procmailrc(5)
  :0c
  $foldername/ # stores in $foldername/new/

  #find out the name of the file we just stored the mail as
  :0
  * LASTFOLDER ?? /\/[^/]+$
  { tail=$MATCH }

  #move the mail to the cur (i.e. read) folder and set satus flags
  TRAP="mv $LASTFOLDER $foldername/cur/$tail:2,S"

  #We still have the original message hanging around
  #so delete it.
  #checking again, so that worst case, if i messed up earlier
  #only spam messages will be deleted.
  :0
  * ^X-Spam-Status: Yes
  /dev/null

  HOST
}