There is an ongoing outage on the primary CPAN mirror. It is possible to work around the issue by using MetaCPAN as a mirror.

parse_xpath

you must indicate which method will be used to parse the received content.

see HTML::Robot::Scrapper::Parser::Default