python-bugzilla adds some state, configuration files and startup overhead without adding any real value today. The usage, to scrape the URL from the review-request page is better done by BeautifulSoup's parsing in urlbug than by the regex matching in bugzilla_bug.

While keeping the interface, where -b 1234567 is actually just a short-cut to an otherwise long url, python-bugzilla could and should be replaced with the urlbug logic giving one dependency less, better startup-time and a shorter(!) manpage.