The first place to look is glob, since it appears that you don't want to recurse down a directory tree. As an example (NOT TESTED)

# insert appropriate shebang line

use strict;
use warnings;

my @files = glob("chat/*.htm"); # use something like [Hh][Tt][Mm] if y
+ou're unsure about the case

@files = grep { -s $_ and -f _ and -r _} @list;

print "Enter search string :";

my $string = <>; # read from STDIN

chomp($string);  # strip off the line feed resulting from <enter>

my $quoted_string = quotemeta($string);  # quote all meta-characters. 
+ Remove if not needed

foreach my $file (@files) {

    my $filehandle;

    unless(open($filehandle, "<", $file)) {
         warn "Could not open $file because $!\n";
         next;
    }

    my @contents = <$filehandle>; # presuming the files are not too la
+rge

    chomp(@contents);
    close($filehandle) or die "Could't close $file because $!\n";


    @contents = grep { /$quoted_string/ } @contents;

    if(@contents){
       print "$string was found in $file:\n\t",join("\n\t",@contents),
+"\n";
   }
}
[download]

Warning! Code is NOT tested, and I won't even guarantee it will compile.

Doubtless there are better ways to do this than my example. One obvious improvement could be to slurp up the entire file into a scalar, not read it into an array, by setting $/ to undef; this is probably faster. Other improvements would be to permit regex.