The oneliner has a BEGIN block to copy @ARGV to the first row of the @AoA then use the current line number $. to choose wich column fill with the current $_ while the current starting column is governed by the $x which is increased every new file read (when eof is reached ARGV is also closed to reset $. ) The END block dump the @aoa in the desired format.

I've changed your sample files (for my sanity) so that file1.txt only contains numbers starting with 1 and so on.

perl -lnE "BEGIN{$aoa[0]=[@ARGV]}$aoa[$.][$x]=$_; if(eof){close ARGV;$
+x++}END{map{say join ',',@$_}@aoa}" file1.txt file2.txt file3.txt

file1.txt,file2.txt,file3.txt
13,22,38
14,25,30
12,27,32
10,23,35
13,28,34
,25,32
,24,32
,,36
[download]

L*