It's generally good form to indicate if you've Crossposted to Stack Overflow and Unix & Linux Stack Exchange to prevent duplication of effort.

The answer I posted on Stack Overflow was:

    #!/usr/bin/env perl
    use strict;
    use warnings;
    use Data::Dumper;
    
    #read file1 into a hash - but invert is it's value => key instead:
    #          'CKEIFJ' => 'cmr03lsp',
    # etc.  
    open( my $file1, '<', "file1.txt" ) or die $!;
    my %file1_content = map { reverse split } <$file1>;
    close($file1);

    print Dumper \%file1_content;
    
    #read file 2 - read keys, store the values. 
    #split _2_ fields, so we keep both numbers as a substring:
    #e.g.:
    #          'cmr03lsp' => '60 70
    #',

    open( my $file2, '<', "file2.txt" ) or die $!;
    my %file2_content = map { split( " ", $_, 2 ) } <$file2>;
    close($file2);

    print Dumper \%file2_content;    
    
    #then iterate file 3, checking if:
    #file1 has a matching 'key' (but inverted - as a value) 
    #file2 has a cross reference. 
    open( my $file3, '<', "file3.txt" ) or die $!;
    while ( my $line = <$file3> ) {
        chomp $line;
        if (    $file1_content{$line}
            and $file2_content{ $file1_content{$line} } )
        {
            print
                "$file1_content{$line} $line $file2_content{$file1_con
+tent{$line}}";
        }
    }
    close($file3);
[download]

Which prints (aside from Dumper diag content):

fji01dde AIDJFMGKG 25 30
cmr03lsp CKEIFJ 60 70
[download]

as requested