You got some good pointers for some one-off code that will likely work just fine for you. None of the answers is particularly robust for real-world URI handling though. So to round it out–

#!/usr/bin/env perl
use strictures;
use URI;
use Path::Tiny;

my @uris = map URI->new($_), <DATA>;
for my $uri ( @uris )
{
    if ( $uri =~ /(\d+)\.htm(l)?$/i )
    {
        print "Naive matched   -> $uri\n";
    }
    else
    {
        print "Naive rejected  -> $uri\n";
    }

    my $file = path( $uri->path )->basename;
    if ( $file =~ /\A[1-9][0-9]{3}\.html?\z/i )
    {
        print "Robust matched  -> $uri\n";
    }
    else
    {
        print "Robust rejected -> $uri\n";
    }
    print "\n";
}

__DATA__
https://moo.cow/queso/1234.HTM
https://www.google.com/search?num=100&q=1234.htm
http://moo.cow/queso/1234.htm?so=what#taco
[download]

Naive matched   -> https://moo.cow/queso/1234.HTM
Robust matched  -> https://moo.cow/queso/1234.HTM

Naive matched   -> https://www.google.com/search?num=100&q=1234.htm
Robust rejected -> https://www.google.com/search?num=100&q=1234.htm

Naive rejected  -> http://moo.cow/queso/1234.htm?so=what#taco
Robust matched  -> http://moo.cow/queso/1234.htm?so=what#taco
[download]

Update: removed superfluous, silly chomp from code.