The script lists the file from the $templates_dir variable, which is setted up in the beginning of the script. For the sake of simplicity, I built the html using the CGI module. Please read the use's to know the dependancies... And, remember it's a CGI, so, unless you're really good in reading HTML, run it in a web serveer...

BTW, there are few output strings in the code, they're in portuguese 'cause I didn't have the courage to translate them... So, if your lost just use a web translator...

#!/usr/bin/perl

use strict;
use Symbol;
use CGI;
use CGI::Untaint;
use OpenOffice::OODoc::Document;
use File::Temp qw(tempfile tempdir);
use utf8;

my $cgi = CGI->new();
my $handler = CGI::Untaint->new($cgi->Vars());
my $templates_dir = '/dir/to/templates/';
my $tmpdir = tempdir();
chdir $tmpdir;

if ($cgi->param('fill')) {
    do_fill();
} elsif ($cgi->param('template')) {
    do_form();
} else {
    do_first_form();
}
exit 0;

sub do_first_form {
    print
      (
       $cgi->header(),
       $cgi->start_html('Selecione o modelo'),
       $cgi->h1('Selecione o modelo de documento'),
       $cgi->start_form(-method => 'GET')
      );
    
    # Obter os modelos disponiveis...
    my $dir_handle = gensym();
    opendir $dir_handle, $templates_dir;
    my @files = grep { /.odt$/ } readdir($dir_handle);
    print $cgi->radio_group
      (
       -name => 'template',
       -values => \@files,
       -linebreak => 'true'
      );
    print
      (
       $cgi->submit(),
       $cgi->end_form(),
       $cgi->hr(),
       $cgi->end_html()
      );
}

sub do_form {
    my $template = $handler->extract( -as_printable => 'template');
    my $document = OpenOffice::OODoc::Document->new(file => $templates
+_dir.$template);
    print
      (
       $cgi->header(),
       $cgi->start_html('Preencha o documento'),
       $cgi->h1("Preenchendo o documento $template"),
       $cgi->start_form(-method => 'GET'),
       $cgi->hidden('template',$template),
       $cgi->hidden('fill','1')
      );
    my %fields = obter_user_fields($document);
    if (%fields) {
        print $cgi->h2('Campos');
        print "<TABLE WIDTH=100%>";
    }
    foreach my $field (sort { $fields{$a}{order} <=> $fields{$b}{order
+} } keys %fields) {
        print
          (
           "<TR><TD>$field</TD><TD width=100%>",
           $cgi->textfield
           (
            -name => 'field_'.$field,
            -default => $fields{$field}{value},
            -style => 'width: 100%'
           ),
           "</TD></TR>".$/
          );
    }
    if (%fields) {
        print "</TABLE>";
    }
    print $cgi->h2('Selecione para remover parágrafos do Documento');
    my @paragraphs = $document->getParagraphList();
    my $count = 0;
    foreach my $para (@paragraphs) {
        my $text = $document->getText($para);
        $text =~ s/\</\&lt;/g;
        $text =~ s/\>/\&gt;/g;
        print "<P><INPUT TYPE=CHECKBOX NAME=para_$count>".$text."</P>"
+;
        $count++;
    }
    print $cgi->submit(), $cgi->end_form(), $cgi->end_html();
}

sub obter_user_fields {
    my $document = shift;
    my @elems = $document->selectElements('//text:user-field-decl');
    my %list;
    my $order = 0;
    foreach my $elem (@elems) {
        my $name = $document->getAttribute($elem,'text:name');
        $list{$name}{value} = $document->getAttribute($elem, 'office:s
+tring-value');
        $list{$name}{order} = $order++;
    }
    return %list;
}

sub do_fill {
    my $template = $handler->extract( -as_printable => 'template');
    my $document = OpenOffice::OODoc::Document->new(file => $templates
+_dir.$template);
    my @params = $cgi->param();
    my (%fields, @paragraphs);
    foreach my $param (@params) {
        if ($param =~ /^field_/) {
            my $field = $param;
            $field =~ s/field_//;
            $fields{$field} = $cgi->param($param);
        } elsif ($param =~ /^para_/) {
            my $para = $param;
            $para =~ s/para_//;
            push @paragraphs, $para;
        }
    }

    my @elems =
      (
       $document->selectElements('//text:user-field-decl'),
       $document->selectElements('//text:user-field-get')
      );

    foreach my $elem (@elems) {
        my $name = $document->getAttribute($elem,'text:name');
        my $value = $fields{$name};
        unless (defined $value) {
            next;
        }
        my $field = $elem->name;
        if ($field =~ /user-field-decl/) {
            $document->setAttributes($elem, 'office:string-value' => $
+value);
        } elsif ($field =~ /user-field-get/) {
            $document->setText($elem,$value);
        }
    }
    
    foreach my $para_num (reverse sort @paragraphs) {
        my $para = $document->getParagraph($para_num);
        $document->removeElement($para);
    }

    # Cuspir o conteúdo do arquivo para STDOUT...
    my ($fh,$filename) = tempfile();
    $document->save($filename);
    close $fh;
    my ($buf,$tempfile);
    open $tempfile, $filename || die $!;
    print $cgi->header
      (
       -type => 'application/octet-stream',
       -attachment => 'preenchido.odt'
      );
    while (read($tempfile,$buf,10)) {
        print $buf;
    }

}
[download]