Notes

Outline

Chapter 7 Phase3:  Gaining Access Using Application and Operating System Attacks

Locating Exploits Packet Storm Security http://packetstorm.securify.com Technotronic Security Information http://www.technotronic.com Security Focus Bugtraq Archives http://www.securityfocus.com

Fig 7.1  Searching Packet Storm for a common vulnerability exploit

Application & Operating System Attacks Stack-based buffer overflow attacks Password attacks Web application attacks

Stack-Based Buffer Overflow Attacks Allows attacker a way to execute arbitrary commands and take control of a vulnerable machine “Smashing the Stack for Fun and Profit” http://packetstorm.securify.com/docs/hack/smashstack.txt Any poorly written application or operating system component could have a stack-based buffer overflow

What is a Stack A data structure that stores important information for processes running on a computer Used to store information associated with function calls on the computer Used to store function call arguments, return instruction pointer, frame pointer, and local variables

Fig 7.2  Sample code with function call

Fig 7.3 A normal stack

Fig 7.4 Buffer Overflow sample program

Fig 7.5  A smashed stack

Contents of a Buffer Overflow Exploit NOP sled Series of “No Operation” instructions Machine language code containing attacker’s commands Return pointer

Buffer Overflow documents Advanced Buffer Overflow Exploit paper http://ohhara.sarang.net/security/adv.txt http://www.blackhad.com/presentations/bh-asia-00/greg/greg-asia-00-stalking.ppt Windows buffer overflow http://www.beavuh.org/dox/win32_oflow.txt eEye’s buffer overflow exploit on Windows NT systems running IIS  http://www.eeye.com/html/advisories/AD19990608.html

Detection of Stack-based overflows by network-based IDS Match signatures associated with NOP sleds Identify typical machine language exploit code to get attackers’ commands executed Look for frequently used return pointers associated with popular buffer overflows

ADMutate Tool used evade IDS detection of buffer overflows http://www.ktwo.ca/security.html exploit code fed into ADMutate which modifies the exploit code while retaining the same ultimate function NOP instruction replaced with other code that functionally does nothing Main part of exploit code contains code to decrypt encrypted instructions Least significant byte of Return Pointer modified

Things Attackers do after Stack is Smashed Force exploit code to spawn a command shell and enter another command to be executed by command shell Shell and command will run under the context of the vulnerable process Installing a backdoor using inetd Backdooring with TFTP and Netcat Shooting back an Xterm

Creating a Backdoor Using Inetd

Backdooring via Netcat Netcat:  A tool used to push a command shell prompt across the network Overflow buffer of victim with command to spawn a shell to download Netcat from attacker’s machine via TFTP and then run Netcat Victim machine runs Netcat configure to execute a shell and push it to the attacker’s machine Attacker’s machine is also running Netcat, but is configured to wait for a connection from victim

Fig 7.6 Placing a backdoor using buffer overflows, TFTP, and Netcat

Shooting back Xterms Useful against networks that block incoming connections but allow outgoing connections Allows attacks to gain command-line access to victim machine victim machine’s configuration need not be modified No additional software needs to be installed on victim machine

Shooting Back Xterms Step-by-Step Attacker configures his own machine to accept incoming X sessions from the target machine via “xhost +victim” Attacker overflows the buffer of vulnerable program on the target machine with  shell command to run the Xterm program and directing the display to the attacker’s machine Commands typed by attacker into Xterm are executed on the victim machine.

Fig 7.7 Getting an Xterm using a buffer overflow

Examples of widely used Exploits IIS Unicode exploit which lets an attacker execute commands on a Windows NT/2000 machine running IIS http://www.wiretrip.net/rft/p/doc.asp?id=57 wu-ftp string input validation problem http://www.kb.cert.org/vuls/id/29823 Rainforest Puppy’s RDS exploit which lets an attacker execute commands on a Windows NT server running IIS http://www.wiretrip.net/rft/p/doc.asp?id=1

Security Mailing Lists BugTraq http://www.securityfocus.com/frames/?content=/forums/bugtraq/intro.html CERT http://www.cert.org/contact_cert/certmaillist.html SANS Newsbite mailing list  http://www.sans.org

Defenses against Stack-Based Buffer Overflow Attacks Keep systems patched Subscribe to security mailing lists Subscribe to vendors’ mailing lists Remove unneeded services from servers Control outgoing traffic such as X

Defenses against Stack-Based Buffer Overflow Attacks (cont.) Configure operating systems with nonexecutable stack Solaris:  add the following to /etc/system file set noexec_user_stack=1 set noexec_user_stack_log=1 Linux: apply a kernel patch http://www.openwall.com/linux/README Windows NT: install SecureStack  http://www.securewave.com/products/securestack/secure_stack.html

Defenses against Stack-Based Buffer Overflow for Software Developers Avoid programming mistakes involving allocation of memory space Check the size of all user input Use automated code-checking tools such as ITS4 (It’s the Software, Stupid – Security Scanner)  http://www.cigital.com/its4/

Password Guessing Attacks Users often choose passwords that are easy to remember, but are also easily guessed default passwords used by vendors left unchanged Database of vendor default passwords  http://security.nerdnet.com

Fig 7.8  An online database of default passwords

Password Guessing through Login Scripting THC-Login Hacker tool http://thc.inferno.tusculum.edu Authforce http://kapheine.hypa.net/authforce/index.php  brute_ssl and brute_web http://packetstrom.security.com/Exploit_Code_archive/brute_ssl.c http://packetstrom.security.com/Exploit_Code_archive/brute_web.c Windows NT password guessing http://packetstorm.securify.com/NT/audit/nt.remotely.crack.nt.passwords.zip Xavier http://www.btinernet.com/~lithiumsoft/ Guessing email passwords using POP3 protocol: Hypnopaedia http://packetstorm.securify.com/Crackers/hypno.zip Other password guessing tools http://packetstorm.securify.com/Crackers

Password Cracking More sophisticated and faster than password guessing through login script Requires access to a file containing user names and encrypted passwords Dictionary attacks Brute force attacks Hybrid dictionary and brute force attacks

Fig 7.9  Password cracking is really just a loop

Password Cracking Tools L0phtCrack, a Windows NT/2000 password cracker http://www.l0pht.com/l0phtcrack John the Ripper, a Unix password cracker http://www.openwall.com/john Crack, a Unix password cracker http://www.users.diron.co.uk/~crypto/ Pandora, a password cracker for Novell  http://www.nmrc.org/pandora PalmCrack, a Windows NT and Unix password cracker that runs on the Palm OS PDA platform  http://www.noncon.org/noncon/download.html

L0phtCrack Tool used to crack Windows NT/2000 passwords Easy to use GUI interface Runs on MS Windows 9x, NT, and 2000 systems Free trial period of 15 days

Cracking Windows NT/2000 Passwords Using L0phtCrack Attacker must get a copy of the encrypted/hashed password representations stored in the SAM database of target machine L0phtCrack includes “pwdump” tool for dumping Windows NT password representation from a local or remote machine across the network Requires administrator privileges on target machine Pwdump3 http://www.ebiz-tech.com/pwdump3/ allows attacker to dump passwords from a SAM database or a Windows 2000 Active Directory

Cracking Windows NT/2000 Passwords Using L0phtCrack (cont.) Boot system from a Linux or DOS floppy disk and retrieve SAM database at %systemroot%\system32\config Since DOS cannot read NTFS partition, attacker can use NTFSDOS program http://packetstorm.securify.com/NT/hack/ntfsdos.zip to access SAM database To access NT and 2000 passwords from Linux boot disk http://home.eunet.no/~pnordahl/ntpasswd/bootdisk.html Use L0phtCrack’s SMB Packet Capture tool to sniff a user’s password off of the network

Fig 7.10  Configuration options for L0phtCrack

Fig 7.11  Successful crack using L0phtCrack

Using L0phtCrack’s Sniffer make the password hash come to you for authentication Send email containing URL file://attacker-pc/sharename/message.html When victim clicks on URL, victim’s machine attempts to mount the share on attacker’s server using a challenge/handshake protocol Password hash is captured by attacker-pc  running L0phtcrack’s integrated sniffing tool Password hash is fed into L0phtcrack to retrieve user’s password

Fig 7.12  Would you trust this email?

Fig 7.13  L0phtCrack’s integrated sniffer captures the challenge/response from the network for cracking

Fig 7.14 Successful crack of sniffed challenge/response

John the Ripper Used to crack Unix and WinNT passwords Runs on Unix, Win9x, NT, and Win2000 systems Automatically detects the encryption algorithm used Quickly generates many permutations for password guesses based on a word list

Fig 7.15 When password shadowing is used, the /etc/passwd file contains no password

Fig 7.16  The corresponding /etc/shadow file contains the encrypted passwords

Retrieving the Encrypted Password File find an exploit that will perform a stack-based buffer overflow of an SUID root program to gain root access Force a process that reads the encrypted password file to generate a core dump (memory dump of a dying process) Crash one instance of a FTP server Use another instance of the FTP server to transfer the core file to look for passwords to crack

Configuring John the Ripper Attacker must feed John with a file that has all user account and password information May need to merge /etc/password and /etc/shadow via “unshadow”

Fig 7.17  Running the unshadow program from John the Ripper

Fig 7.18  Running John the Ripper to crack passwords

Defenses against Password-Cracking Attacks Do not select passwords that can be easily guessed by an automated tool Do not use dictionary terms Change passwords at specified intervals Know how to create a good password Use first letters of each word from a memorable phrase, mixing in numbers and special characters Use password filtering software to prevent users from choosing easily guessed  passwords Use one-time password tokens or smart cards Use 2 or 3 factor authentication

Password Filtering Software Unix platform Npasswd ftp.cc.utexas.edu/pub/npasswd Passwd+ ftp.dartmouth.edu/pub/security Windows NT Passprop, available in MS WinNT Resource Kit Passfilt.dll included in Service Pack 2 Password Guardian www.georgiasoftworks.com Strongpass http://ntsecurity.nu/toolbox Fast Lane http://www.fastlanetech.com

Web Application Attacks Can be conducted even if the Web server uses Secure Sockets Layer (SSL) SSL used to authenticate the Web server to the browser SSL used to prevent an attacker from intercepting traffic SSL can be used to authenticate the client with client-side certificates Web attacks can occur over SSL-encrypted connection Account harvesting Undermining session tracking SQL Piggybacking

Account Harvesting Technique used to determine legitimate userIDs and even passwords of a vulnerable application Targets the authentication process when application requests a userID and password Works against applications that have a different error message for users who type in an incorrect userID

Fig 7.19  Mock Bank’s error message when a user types an invalid userID

Fig 7.20  Mock Bank’s error message when a user types a valid userID, but the wrong password

Account Harvesting Defenses Make sure that error message is the same  when a user types in an incorrect userID or password

Web Application Session Tracking Most Web application generate a session ID to track the user’s session. Session ID is passed back and forth across the HTTP or HTTPS connection when client browses web pages, enters data into forms, or conducting transactions Session ID allows the Web application to maintain the state of a session with a user Session ID is independent of the SSL connection Session ID is Application-level data

Implementing Session IDs in Web Applications URL session tracking Session ID is written directly on browser’s location line Hidden form elements Hidden Session ID element put into the HTML form Session ID can be seen by user by viewing HTML source code <INPUT TYPE=“HIDDEN” MAME=“Session” VALUE=“22343”> Cookies Most widely used session-tracking method Cookie is an HTTP field that the browser stores on behalf of a Web server, containing info such as user preference and session ID Per-session cookie is stored in browser’s memory Persistent cookie is written to the local file system of client

Fig 7.21  Session tracking using the URL

Attacking Session Tracking Mechanisms Attacker changes his session ID to a value assigned to another user Application thinks that attacker is the other user

Fig 7.22  Editing persistent cookies to modify a session ID using notepad

Achilles Tool used to edit per-session cookies www.digizen-securitycom A Web proxy Attacker’s browser configured to send all HTTP and HTTPS data to Achilles Web browser and proxy can run on same or different machines Archilles allows attacker to edit all HTTP/HTTPS fields, per-session and persistent cookies, hidden form elements, and URLs. Supports HTTPS connections one SSL connection set up between browser and Achilles Another SSL connection set up between Achilles and Web server

Fig 7.24  The Achilles screen

Fig 7.25  Handling HTTPS with Achilles

Defending against Web Application Session-Tracking Attacks Digitally sign or hash session-tracking information Encrypt information in the URL, hidden form element, or cookie Make sure that your session IDs are long enough to prevent accidental collision Apply a timestamp within the session ID variable and encrypt it Allow users to terminate their sessions via a logout button which will invalidate the session ID Scan your web site via  AppScan http://www.sanctuminc.com

SQL Piggybacking Attacker may can extend an application’s SQL statement  to extract or update information that the attacker is not authorized to access “How I Hacked Packetstorm”  http://www.wiretrip.net/rfp/p/doc.asp?id=42 Attacker will explore how the Web application interacts with the back-end database by finding a user-supplied input string that will be part of a database query

Fig 7.26 Figuring out how the Web application interacts with a database

Fig 7.27  The location line contains the account number searched for

Fig 7.28 A very useful error message

SQL Statement used by application

Fig 7.29  Gaining unauthorized access with SQL piggybacking

Defenses against Piggybacking SQL Commands Web application must be programmed to carefully filter user-supplied data Potentially damaging characters (such as ‘ ”   ` ; * % _  ) should be filtered at server side World Wide Web Security FAQ  http://www.w3.org/Security/Faq/www-security-faq.html