Polymorphic code

Most anti-virus software and intrusion detection systems (IDS) attempt to locate malicious code by searching through computer files and data packets sent over a computer network. If the security software finds patterns that correspond to known computer viruses or worms, it takes appropriate steps to neutralize the threat. Polymorphic algorithms make it difficult for such software to recognize the offending code because it constantly mutates.

Malicious programmers have sought to protect their encrypted code from this virus-scanning strategy by rewriting the unencrypted decryption engine (and the resulting encrypted payload) each time the virus or worm is propagated. Anti-virus software uses sophisticated pattern analysis to find underlying patterns within the different mutations of the decryption engine, in hopes of reliably detecting such malware.

Emulation may be used to defeat polymorphic obfuscation by letting the malware demangle itself in a virtual environment before utilizing other methods, such as traditional signature scanning. Such a virtual environment is sometimes called a sandbox. Polymorphism does not protect the virus against such emulation if the decrypted payload remains the same regardless of variation in the decryption algorithm. Metamorphic code techniques may be used to complicate detection further, as the virus may execute without ever having identifiable code blocks in memory that remains constant from infection to infection.

The first known polymorphic virus was written by Mark Washburn. The virus, called 1260, was written in 1990. A better-known polymorphic virus was created in 1992 by the hacker Dark Avenger (a pseudonym) as a means of avoiding pattern recognition from antivirus software. A common and very virulent polymorphic virus is the file infecter Virut.

This example is not really a polymorphic code but will serve as an introduction to the world of encryption via the XOR operator. For example, in an algorithm using the variables A and B but not the variable C, there could be a large amount of code that changes C, and it would have no effect on the algorithm itself, allowing it to be changed endlessly and without heed as to what the final product will be.

    much encrypted code
    ...
Decryption_Code:
    C = C + 1
    A = Encrypted
Loop:
    B = *A
    C = 3214 * A
    B = B XOR CryptoKey
    *A = B
    C = 1
    C = A + B
    A = A + 1
    GOTO Loop IF NOT A = Decryption_Code
    C = C^2
    GOTO Encrypted
 CryptoKey:
    some_random_number

The encrypted code is the payload. To make different versions of the code, in each copy the garbage lines which manipulate C will change. The code inside "Encrypted" ("lots of encrypted code") can search the code between Decryption_Code and CryptoKey and each algorithm for new code that does the same thing. Usually, the coder uses a zero key (for example; A xor 0 = A) for the first generation of the virus, making it easier for the coder because with this key the code is not encrypted. The coder then implements an incremental key algorithm or a random one.

Polymorphic code can be also used to generate encryption algorithm. This code was generated by the online service StringEncrypt.[3] It takes the string or a file content and encrypts it with random encryption commands and generates polymorphic decryption code in one of the many supported programming languages:

     // encrypted with https://www.stringencrypt.com (v1.1.0) [C/C++]
     // szLabel = "Wikipedia"
     wchar_t szLabel[10] = { 0xB1A8, 0xB12E, 0xB0B4, 0xB03C, 0x33B9, 0xB30C, 0x3295, 0xB260,
                             0xB5E5, 0x35A2 };
 
     for (unsigned int tUTuj = 0, KRspk = 0; tUTuj < 10; tUTuj++)
     {
             KRspk = szLabel[tUTuj];
             KRspk ^= 0x2622;
             KRspk = ~KRspk;
             KRspk --;
             KRspk += tUTuj;
             KRspk = (((KRspk & 0xFFFF) >> 3) | (KRspk << 13)) & 0xFFFF;
             KRspk += tUTuj;
             KRspk --;
             KRspk = ((KRspk << 8) | ( (KRspk & 0xFFFF) >> 8)) & 0xFFFF;
             KRspk ^= 0xE702;
             KRspk = ((KRspk << 4) | ( (KRspk & 0xFFFF) >> 12)) & 0xFFFF;
             KRspk ^= tUTuj;
             KRspk ++;
             KRspk = (((KRspk & 0xFFFF) >> 8) | (KRspk << 8)) & 0xFFFF;
             KRspk = ~KRspk;
             szLabel[tUTuj] = KRspk;
     }
 
     wprintf(szLabel);

As you can see in this C++ example, the string was encrypted and each character was stored in encrypted form using UNICODE widechar format. Different encryption commands were used like bitwise XOR, NOT, addition, subtraction, bit rotations. Everything is randomized, encryption keys, bit rotation counters and encryption commands order as well. Output code can be generated in C/C++, C#, Java, JavaScript, Python, Ruby, Haskell, MASM, FASM and AutoIt. Thanks to the randomization the generated algorithm is different every time. It's not possible to write generic decryption tools and the compiled code with polymorphic encryption code has to be analyzed each time it's re-encrypted.

• Timeline of notable computer viruses and worms
• Metamorphic code
• Self-modifying code
• Alphanumeric code
• Shellcode
• Software cracking
• Security cracking
• Obfuscated code
• Oligomorphic code

• Spinellis, Diomidis; Reliable identification of bounded-length viruses is NP-complete, IEEE Transactions on Information Theory, 49(1):280–284, January 2003. doi:10.1109/TIT.2002.806137