Due to the immutability of smart contracts, checking their security before deploying them in any Blockchain is one of the most critical steps in their development lifecycle. However, this process ...
Honeypot programs are one of the best tools that security researchers have ever made to study the new or unknown hacking techniques used by attackers. Therefore, using honeypots in smart contract could be a very good idea to study those attacks. So what is honeypot in smart contract?
Honeypots in the Blockchain industry is an intentionally vulnerable smart contract that was made to push attackers to exploit its vulnerability. The idea is to convince attackers or even simple users to send a small portion of cryptocurrency to the contract to exploit it, then lock those ethers in the contract.
In this blog post, you are going to see some examples of those honeypots with a detailed technical explanation of how they work. So if you are interested to learn more about this subject just keep reading and leave a comment at the end.
A honeypot is a smart contract that purports to leak cash to an arbitrary user due to a clear vulnerability in its code in exchange for extra payments from that user. The monies donated by the user to the vulnerable contract get then locked in the contract and only the honeypot designer or attacker will be able to recover them.
The concept of a honeypot is well known in the field of network security and was used for years by security research. The main objective of using them was to identify new or unknown exploits or techniques already used in the wild. In addition, Honeypots were used to identify zero-day vulnerabilities and report them to vendors. This technique was basically designed to trap black hat hackers and learn from them.
However, with the rise of Blockchain technology and the smart contract concept. Black hat hackers started to use this concept to trap users both with good or bad intentions. The idea is simple, the honeypot designer creates a smart contract and puts a clear vulnerability in it. Then hid a malicious code in its smart contract or between its transactions to block the right execution of the withdraw function. Then he deploys the contract and waits for other users to get into the trap.
What actually makes this concept even more dangerous in the context of blockchain is that implementing a honeypot is not really difficult and does not require advanced skills. In fact, any user can implement a honeypot in the blockchain, all it needs is the actual fees to deploy such a contract in the blockchain.
In fact, in the blockchain, the word “attacker” could be given to both the one who deploys the smart contract honeypot and the one trying to exploit it (depending on his intention). Therefore, in the following sections of this blog post, we will use the word “deployer” to the one who implements the honeypot and “user” to the one trying to exploit that smart contract.
Honeypots in smart contract can be divided into 3 main categories depending on the used techniques:
The main idea of honeypot in the network context is to supervise an intentionally vulnerable component to see how it can be exploited by hackers. However, in smart contract the main idea is to hide a behavior from users and trick them to send ether to gain more due to the vulnerability exploitation.
Therefore, what actually defines each smart contract honeypot category is the used technique to hide that information from users.
The first category of smart contract honeypot is based on the way the EVM instruction is executed. It is true that the EVM follow an exact set of rules, however, some instruction requires a very good experience with the way EVM works to be able to detect the honeypot otherwise the user could easily be fooled.
The second category of smart contract honeypot is related to the solidity compiler. In other words, the smart contract honeypot builder should have a good experience with smart contract development and a deep understanding of how Solidity compiler would work. For example, the way inherence is managed by each version of the solidity compiler, or when overwriting variables or parameters would happen.
The third category of smart contract honeypot is based on hiding things from the users. Most users that try to exploit a program look for the easier way to do so (quick wins). Therefore, they may not take the time to analyze all parts of the vulnerable smart contract. This user behavior leads to locking his money in the smart contract. In this blog post, we are going to discuss 4 techniques used by deployers to hide an internal behavior from the users and therefore fool the user.
The EVM-based smart contract honeypots have only one subtype called balance disorder. I think the best way to understand how this type of smart contract honeypots works, is by example. So take a look at the following example:
This example is taken from the following contract: https://etherscan.io/address/0x8b3e6e910dfd6b406f9f15962b3656e799f60d2b#code
A quick look at this function from a user, he can easily understand that if he sends while calling this function more than what the contract balance actually has, then everything in the contract plus what he sends will be sent back to him. Which is obviously a good deal.
However, what a user could miss in this quick analysis of the smart contract is that the contract balance will be incremented as soon as the function of the call is performed by the user. This means that the msg.value will always be lower than the contract balance no matter what you do. Therefore, the condition will never be true and the contract will be locked in this contract.
Another example of the balance disorder type of honeypot could be found here:
By visiting this link you can see that there is no source code out there. So there are two ways to analyze this contract. The first one and the most difficult is to get the bytecode of this smart contract and then try to understand and reverse engineer it. Or the second way is to try to decompile it using different tools available to get an intermediate and easy-to-understand source code.
I personally used the second technique to accelerate the analysis and simply used the Etherscan default decompile. In the smart contract you want to decompile you can click here:
And wait for a moment about 30 seconds to get the source code.
By taking a look at the source code, and especially at the “multiplicate” function you can now easily see the same logic as the previously explained example.
The condition in line 24 will never be verified and the money will be stuck in the contract.
As I said, this category of smart contract honeypots is based on some deep knowledge about how the Solidity compiler works. In the following subsection, I will give you 4 techniques that are used to build this kind of smart contract honeypots. However, other unknown techniques might be used in the wild, and I will do my best to update this blog post whenever I found a new one. Please comment below and tell me if you know a technique that was not noted in this blog post.
One of the most confusing systems in solidity language or even in other programming languages is inheritance. A lot of hidden aspects in this concept could be used by deployer to fool the users and work contrary to what is expected.
In solidity language, a smart contract can implement the inheritance concept by using the word “is” followed by the different smart contract that this one wants to inherit their source code. Then only one smart contract is created and the source code from the other contracts is copied into it.
To better understand how such a mechanism could be exploited to create honeypots please take a look at the following examples:
You can find this contract here: https://etherscan.io/address/0xd3bd3c8fb11429b0deee5301e72b66fba29782c0#code
If you take a look at this contract source code, you can easily notice that it has an obvious vulnerability related to access control. The function setup allows a user to change the owner of this contract without checking if he is the actual owner. Therefore, the user would be able to execute the withdraw function to get the money.
However, this analysis assumes that the isOwner() function inherited from the Ownable contract is going to check the local variable Owner.
Unfortunately, this is not what will actually happen. The inheritance creates a different variable for each contract even if they have the same name. The variable Ownable.Owner is totally different than the ICO.Owner. Therefore, when the user will call the setup() function, this one will change the value of ICO.Owner and not Ownable.Owner. This means that the result of the isOwner() will remain the same.
Another example of this same type of solidity compiler-based honeypot can be found here. The same logic applies to this smart contract. The Owner variable will not change by calling the setup() function.
Another tricky behavior in solidity compiler that may not be very easy to discover is the skip empty string literal. The skip empty string literal problem happens in solidity when a function is called with an empty string as a parameter. This is a known bug in solidity compilers before 0.4.13 here is a reference for it.
The encoder skips the empty string literal “” when used as a parameter in a function call. As a result, the encoding of all subsequent arguments is moved left by 32 bytes, causing the function call data to be malformed.
This kind of honeypot could be easily detected, by just looking at the solidity compiler version and then scrolling down the source code to see if there is any use of the empty string in a function call. However, a knowledge of this bug is required to detect the problem in the smart contract.
Here is a simple example of this honeypot:
Check the following smart contract: https://etherscan.io/address/0x2b990227344300aded3a072b3bfb9878b209da0a#code
The source code is a little bit long so I will put just the most important functions:
In the divest() function line 83, the external function call to loggedTransfer() with the empty string will result in shifting the parameters by 32 bytes which leads to replacing the target address from msg.sender to the owner address. Therefore, the user will send the money to the owner of the contract and not his own address. This simply means that the user will never be able to retrieve the money he sent to this smart contract.
The Solidity compiler offers a nice feature that helps developers declare a variable without knowing exactly what type it would be. This could be made by creating a variable with the keyword “var” and the compiler will deduce what type is better for that result. However, this technique may cause a problem called type deduction overflow.
This problem could be used in a smart contract honeypot to cause a revert and then lock the money on the contract. To better illustrate this problem please take a look at the following source code:
You can check the whole code here:
In this contract the Double() function allow a user to double his money by first sending at least more than one ether and then looping to create the value of the ethers that will be sent to the user. This seems to be a nice and easy smart contract to exploit.
However, this contract loop will never reach even half of the value sent by the user. The reason behind this is the way the variable “i” is declared. The “var” keyword, will create a variable with a type of uint8 due to the 0 value affected to it in the beginning. The code should loop till it gets to msg.value which is a uint256 and the value would be more than 1 with 18 digits. However, the size of the “i” variable can only reach 255 then once incremented will get back to 0. Therefore, the loop will end and all that the user will receive is 255 wei.
The uninitialized structure is a common problem in solidity and could be seen both as a vulnerability and as a way to trick users. In this blog post, I am going to discuss the tricky part of this problem. However, if you want me to discuss how this could be a vulnerability, please comment below and I will be happy to make a blog post about it.
An uninitialized structure problem happens when a structure variable is not initialized at the moment of its creation. When a structure variable is not initialized in the same line as its creation with the keyword “new”, the solidity compiler point that variable to the first slot of the smart contract. This simply means the variable will be pointing to the first variable of the smart contract. Once the developer starts affecting values to the structure variable, the first element value of the structure will overwrite the first variable value.
This concept is used by smart contract honeypots deployer to trick users to send money to exploit an obvious vulnerability in it.
Here is an example of such a honeypot:
This contract asks the user to guess a number while betting with some of his money. The secret value that a user is going to guess is stored in the first slot of the smart contract. For a quick analysis of this contract, the user would assume that the contract is vulnerable as even private variables could be seen in the Blockchain.
However, once the user will call the play() function and send money to it, the function will create a structure “game” in line 51 without correctly initializing it. This means that this structure variable will point to the first slot (variable secretNumber). In addition, the game.player will be the variable that will overwrite the secretNumber variable. Therefore, the user “would not” will not be able to correctly guess the number.
Actually, in this example, the honeypot could be bypassed to retrieve the money. If you take a look at the value affected to the game.player variable that overwrite the secretNumber. You will see that it is simply the sender’s address. Therefore, the value the user should send, is simply his address converted to decimals.
All the smart contracts that we have seen until now, exploit a solidity language gap of knowledge in the user. However, in this section of this blog post, the deployer exploits some features related to etherscan platform to hide some important information that may trick users.
The Etherscan platform helps developers and any Ethereum Blockchain user to debug his smart contract or track his transactions. Therefore, the platform display user’s transaction and internal messages that are performed by smart contracts. However, one of the features of Etherscan is that it does not show internal messages with an empty value.
Therefore, smart contract honeypot deployer exploit this feature to trick users and change the smart contract behavior. Here is an example to better understand this concept:
Check the following smart contract: https://etherscan.io/address/0x8bbf2d91e3c601df2c71c4ee98e87351922f8aa7#code
This contract might be used as a honeypot, as the user could be fooled by the initial value of the variable passHasBeenSet. By checking the Etherscan data he would not be able to see any transaction that has changed the value of passHasBeenSet. Therefore, he would assume that the value didn’t change and attempt to exploit the contract.
To do that, the user would try to exploit the contract by sending more than one ether to the contract using the GetGift() after setting the hashPass using SetPass() function.
However, the passHasBeenSet variable might be already changed by another contract and that would not be seen in the etherscan platform.
This technique is built upon showing a source code for a contract that is not actually the one used by the contract. For example, the deployer could build a contract that requires another library and that that library address is initialized during the deployment of the contract or by calling a specific function.
At this stage, there is nothing that holds the deployer from using another contract address that is totally different than the one that the source code is displayed in Etherscan.
Unfortunately, this really a tricky honeypot and a really difficult technique to discover from a user. I mean the user should verify the addresses of the deployed contract and the different transactions and data passed to the contract to be able to find this issue. Moreover, even if the user tries to test this smart contract in a different contract, he will use the smart contract code displayed by the attacker and he will see a normal behavior. Which makes it even more difficult to find the issue.
Here is an example of such a honeypot, try to take a look at it and see what makes this smart contract a honeypot:
If you want to go even deeper into this subject, here are some very good resources:
Written by: Z. Oualid
I am a Cyber Security Expert, I have worked with many companies around the globe to secure their applications and their networks. I am certified OSCP and OSCE which are the most recognized and hard technical certifications in the industry of cybersecurity. I am also a Certifed Ethical hacker (CEH). I hope you enjoy my articles :).
In the intricate landscape of cybersecurity, understanding the concept of malware artifacts is paramount for safeguarding digital environments. All malwares has their own digitale signature that is represented by their [...]