Recently, CF began blocking the default urllib2 python user agent (at least for me). This means that requests made using this script fail with a "Forbidden 403" error.
Without looking to see if CF has a preferred way of interacting, I'll note that manually changing the user agent allows this simple script to continue working.