Though it seems to work, I have a hunch this isn't ideal.

Of course, the best way is to use a proper parser. In this case, the original Markdown is a Perl script itself, and the regex extracted from it looks like the following (modified slightly to put it into qr// form). Of course, now there are tons of different Markdown variations and parsers, and probably several that are more robust than this "just a regex" parser, but as long as your Markdown input is simple enough, it should probably be ok. <update> There are caveats to this approach, though - for example, the following regex will also operate on code blocks! As usual, more representative sample data will result in more accurate solutions :-) </update>

my $g_nested_brackets;
$g_nested_brackets = qr{
    (?>                             # Atomic matching
       [^\[\]]+                     # Anything other than brackets
     | 
       \[
         (??{ $g_nested_brackets }) # Recursive set of nested brackets
       \]
    )*
}x;
my $anchors = qr{
    (                # wrap whole match in $1
      \[
        ($g_nested_brackets)    # link text = $2
      \]
      \(            # literal paren
        [ \t]*
        <?(.*?)>?   # href = $3
        [ \t]*
        (           # $4
          (['"])    # quote char = $5
          (.*?)     # Title = $6
          \5        # matching quote
        )?          # title is optional
      \)
    )
}xs;
[download]

I've taken this regex and modified it to modernize it a bit and only capture the things we're interested in:

use warnings;
use strict;

my $anchors = qr{
    (?(DEFINE) (?<nested_brackets>
        (?> [^\[\]]+ | \[ (?&nested_brackets) \] )*
    ) )
    \[ (?<text> (?&nested_brackets) ) \]
    \( (?<link>
        [ \t]* <? .*? >? [ \t]*
        (?: (?<titlequote>['"]) .*? \k<titlequote> )?
    ) \)
}xs;

my $input = <<'END';
blah blah [click me](click me) more stuff
blah [link here](link here) blah blah
END
my $expect = <<'END';
blah blah [click me](/click-me) more stuff
blah [link here](/link-here) blah blah
END
(my $output = $input) =~
    s{$anchors}{
        my ($t, $l) = @+{qw/ text link /};
        $l =~ s/\s+/-/g;
        "[$t](/$l)"
    }ge;

use Test::More tests=>1;
is $output, $expect;
[download]

Holy cow, there's some really advanced regex stuff going on here I've never seen before. I will study this closely and pick up some new tricks. Thanks.

I thought about researching a markdown parser but I figured it was easier, for now, just to roll my own before going down a big wormhole.

$PM = "Perl Monk's";
$MCF = "Most Clueless Friar Abbot Bishop Pontiff Deacon Curate Priest Vicar";
$nysus = $PM . ' ' . $MCF;
Click here if you love Perl Monks

Call a function in the substitution part (activated by /e modifier) which returns your transformation.

s/PATTERN/replace($1,$2)/eg

Not sure if you even need to pass the matchgroups $1,... into the sub, but it's certainly more readable.

Edit

Pattern should match the links [$1]($2) (pseudo syntax).

I recommend using the /x flag for readability.

The transformations of $2 should happen inside the replace() sub

HTH!

Very cool. Didn't know that was even possible (well probably read about it once and then forgot). Thanks.

$PM = "Perl Monk's";
$MCF = "Most Clueless Friar Abbot Bishop Pontiff Deacon Curate Priest Vicar";
$nysus = $PM . ' ' . $MCF;
Click here if you love Perl Monks

#!/usr/bin/perl

use strict; # https://perlmonks.org/?node_id=11119206
use warnings;

my $text = "blah blah [click me](click me) more stuff\nblah [link here
+](link here) blah blah";

s!]\(\K.*?\)! '/' . $& =~ tr~ ~-~r !ge for $text;

print "$text\n";
[download]