There's also PPI, which does a relatively good job of parsing Perl.

With PPI it should certainly be possible to identify and check that subset of your returns and function calls which take some regular, identifiable form. I'm not (yet) aware of existing code that already does this, but I could be missing something. Maybe one of the many Perl::Critic policies could provide a starting point...

use PPI; 
my $doc = PPI::Document->new( \<<'ENDPERL' );
sub foo { return 1, 2, 3 }
my ($x, $y, $z) = foo();
ENDPERL
use PPI::Dumper;
PPI::Dumper->new( $doc, whitespace=>0 )->print;
__END__
PPI::Document
  PPI::Statement::Sub
    PPI::Token::Word      'sub'
    PPI::Token::Word      'foo'
    PPI::Structure::Block      { ... }
      PPI::Statement::Break
        PPI::Token::Word      'return'
        PPI::Token::Number      '1'
        PPI::Token::Operator      ','
        PPI::Token::Number      '2'
        PPI::Token::Operator      ','
        PPI::Token::Number      '3'
  PPI::Statement::Variable
    PPI::Token::Word      'my'
    PPI::Structure::List      ( ... )
      PPI::Statement::Expression
        PPI::Token::Symbol      '$x'
        PPI::Token::Operator      ','
        PPI::Token::Symbol      '$y'
        PPI::Token::Operator      ','
        PPI::Token::Symbol      '$z'
    PPI::Token::Operator      '='
    PPI::Token::Word      'foo'
    PPI::Structure::List      ( ... )
    PPI::Token::Structure      ';'
[download]