#!/usr/bin/perl
use warnings;
use strict;

use Data::Dumper;

my $string = 'x:q,y:u,z:n';
my %hash = split /[:,]/, $string;
print Dumper \%hash;

$string = '[x:q,y:u,z:n]';
%hash = split /[:,]/, substr $string, 1, -1;
print Dumper \%hash;
[download]

Both given answers are fine, but this smells like records embedded in CSV, where both answers will soon fail. What if x is allowed to be «This, my friend, is why : is allowed in "text" to fail»?

update: additional remarks:

In case that cannot happen, stick to the earlier answer. They are fast and simple. In case it /can/ happen, you need a double CSV parser. The first to split on , the second on ::

my $outer = Text::CSV_XS->new ({ binary => 1, auto_diag => 1 });
my $inner = Text::CSV_XS->new ({ binary => 1, auto_diag => 1, sep_char
+ => ":" });
my %hash;
while (my $row = $outer->getline ($fh)) {
    foreach (@$row) {
        $inner->parse ($_);
        my @f = $inner->fields;
        if (@f >= 2) { # assuming key:value:1 will result in $hash{key
+} = "value:1"
            my $key = shift @f;
            $hash{$key} = join ":" => @f;
            }
        else {
            warn "Probably not a key-value pair '$_'";
            }
        }
    }
[download]

which is likely to get more and more complicated as you proceed.

Or this:

use warnings;
use strict;
use Data::Dumper;

my %h = map $_ => split/:|,/ => 'x:q,y:u,z:n';

{ 
 local $Data::Dumper::Sortkeys = 1;  ## Update
 print Dumper \%h;
}
[download]