You are too focused on the immediate solution and haven't provided enough context for us to test suggestions against. A bit of overview of the big picture problem can save a lot of iterations while we slowly leach enough context out of you to be helpful. Something like the following problem description could help a lot:
I have a schedule file with named lists of tasks. The task lists include actors involved in the tasks. I also have a file containing a list of actors I'm interested in, and a definitions file matching actors to the jobs they do. I want to generate an output list indicating which actors are used and, if they are used, which schedule they appear in.

Given that and assuming the sample data from my previous example the following fits the bill:

.
.
.
open my $schdIn, '<', \$scheduleData or die "Can't open 'Schedule': $!
+";
my %actorSchedules;
my $scheduleName;

while (defined(my $line = <$schdIn>)) {
    if ($line =~ /SCHEDULE\s+(\S+)/) {
        $scheduleName = $1;
        next;
    }

    next if $line !~ /(\S+)#(\S+)/;
    ++$actorSchedules{$1}{$scheduleName};
}

open my $agentsIn, '<', \$agentData or die "Can't open 'Agents': $!";

my $agentsList  = join '|', map {chomp; qr/\Q$_\E/} <$agentsIn>;
my $agentsMatch = qr|\b($agentsList)\b|;
my $wantedSchedule = 'Wanted';

open my $defsIn, '<', \$defsData or die "Can't open 'Definitions': $!"
+;

while (defined(my $defLine = <$defsIn>)) {
    chomp $defLine;

    next if $defLine !~ /$agentsMatch#/;

    my $lineout = $defLine;
    my $tail   = $actorSchedules{$1}           ? 'Yes'    : 'No';
    my @schedules  = keys %{$actorSchedules{$1}};

    $tail = join ' | ', $tail, @schedules;
    print "$defLine | $tail\n";
}
[download]

Prints:

Mindy#mu4 | Yes | Wanted
Orlon#mu5 | No
Nancy#mu6 | Yes | Unwanted
Harry#mu7 | No
[download]

Of course I don't know just what your big picture goal is so it's likely I haven't hit the mark exactly.

I hear what you're saying, I just am not sure which part I'm not adequately explaining, perhaps I can example it better without actual code. If you see something wrong in the logic let me know

Step 1 starts with a list of agents, the source data we've already assigned $agentsIn which reads from $agentsData

Harry
Mindy
Nancy
Orlon
[download]

Step 2 looks over every agent in the previous list, and prints every matching line based on the regex, the full line because the full line contains information we need out of schedulesData later. In this example it should return Mindy#mu4, Orlon#mu5, Nancy#mu6, Harry#mu7, Mindy#mu8 because Ted isn't in $agentsData, but the rest are.

Ted#mu3
 SCRIPTNAME "doit.cmd"
 DESCRIPTION "mu2"
Mindy#mu4
 SCRIPTNAME "doit.cmd"
 DESCRIPTION "mu2"
Orlon#mu5
 SCRIPTNAME "doit.cmd"
 DESCRIPTION "SOMETIMES HAS AGENTNAME Orlon"
Nancy#mu6
 SCRIPTNAME "doit.cmd"
 DESCRIPTION "mu2"
Harry#mu7
 SCRIPTNAME "SOMETIMES HAS AGENTNAME Harry"
 DESCRIPTION "mu2"
Mindy#mu8
 SCRIPTNAME "doit.cmd"
 DESCRIPTION "mu2"
[download]

Step 3 wants to know if the 5 lines we just found are found in the $schedulesData we're pulling from $schedulesIn file.
Mindy#mu4 is in the list below, return 'yes' BUT also save this item for next step
Orlon#mu5 is NOT in the list below, return 'no'.
Nancy#mu3 is in the list below, return 'yes' BUT also save this item for next step
Harry#mu7 is NOT in the list below, return 'no'.
Mindy#mu8 is in the list below, return 'yes' BUT also save this item for next step


SCHEDULE magicName1
mu1
mu2
Nancy#mu9
 mu4
 mu2
END
SCHEDULE magicName2
mu1
mu2
Mindy#mu4
 mu4
 mu2
END
 SCHEDULE magicName3
mu1
mu2
Nancy#mu3
 mu4
 mu2
END
SCHEDULE magicName4
mu1
mu2
Mindy#mu8
 mu4
 mu2
END
[download]

Step 4 is the tricky part I've tried a few examples and different ways to explain but thus far am not getting it across. There should be 3 matches in the above, the final step is figuring out what the closest magicName in the file above the matches are. This was the functionality in my second script, my problem was figuring out how to hit on the match and look upward to the previously last match of the other regex while in the previous loop. This is what I meant with $wantedSchedule can't be static, its found by storing all hits on SCHEDULE(.*) to figure out what the magicName is and then using the agent name to determine which one applies to that record

use Mindy#mu4 to find magicName2
use Nancy#mu3 to find magicName3
use Mindy#mu8 to find magicName4


The final output would be like this

Harry#mu7 | No | No
Mindy#mu4 | Yes | magicName2
Mindy#mu8 | Yes | magicName4
Nancy#mu6 | No | No
Nancy#mu3 | Yes  | magicName3
Orlon#mu5 | No | No
[download]

The second column of Yes/No isn't really needed in the final solution, it was used as an intermediary step to get me toward the final solution which turned out to be harder than expected. If its yes it means there will be a record in the last step, if its no then the last step will also be no, but I wanted to start simple by determining first if a match is in the file at all then hit it with a more complex function to find the match and tell me the closest magicName above. magicName is not always the same distance above a match, so there's no good way to figure it out without storing all the magicNames and just printing the last one stored when I reach the input match because the match we search for will always be lower in the record than magicName we actually want to know.