Understanding the Examples: In the "Examples" section, we delve into practical instances of front-running vulnerabilities unearthed from actual bug bounty reports, Code Arena analyses, and more. By studying these real-life cases, you'll gain a deeper, more tangible understanding of front running and its potential impact, enhancing your auditing skills in a practical and impactful way. Example 4 and 5 are my personal findings related to front running.
Based on our previous discussion, see if you can find the Front running vulnerability in the code bellow:
In this example Adam creates a guessing game. You win 10 ether if you can find the correct string that hashes to the target hash and you are the first person to submit it. Let's see how this contract is vulnerable to front running
contract FindThisHash {
bytes32 public constant hash =
0x564ccaf7594d66b1eaaea24fe01f0585bf52ee70852af4eac0cc4b04711cd0e2;
constructor() payable {}
function solve(string memory solution) public {
require(hash == keccak256(abi.encodePacked(solution)), "Incorrect answer");
(bool sent, ) = msg.sender.call{value: 10 ether}("");
require(sent, "Failed to send Ether");
}
}
Did you find it? Great! if not, consider this scenario:
Adam deploys FindThisHash with 10 Ether.
Bob finds the correct string that will hash to the target hash. ("Ethereum")
Bob calls solve("Ethereum") with gas price set to 15 gwei.
Eric is watching the transaction pool for the answer to be submitted.
Eric sees Bob's answer and calls solve("Ethereum") with a higher gas price than Bob (100 gwei).
Eric transaction was mined before Bob's transaction. Eric won the reward of 10 ether
What happened?
Transactions take some time before they are mined. Transactions not yet mined are put in the transaction pool. Transactions with higher gas price are typically mined first. An attacker can get the answer from the transaction pool, send a transaction with a higher gas price so that their transaction will be included in a block before the original.
Example 2: Real life Cod4rena Bug Find With Payout (Medium Risk Vulnerability)
See if you can find the Front running vulnerability in the code bellow!
/// @dev Create a new vault
function build(address owner, bytes12 vaultId, bytes6 seriesId, bytes6 ilkId)
external
auth
returns(DataTypes.Vault memory vault)
{
require (vaultId != bytes12(0), "Vault id is zero");
require (vaults[vaultId].seriesId == bytes6(0), "Vault already exists"); // Series can't take bytes6(0) as their id
require (ilks[seriesId][ilkId] == true, "Ilk not added to series");
vault = DataTypes.Vault({
owner: owner,
seriesId: seriesId,
ilkId: ilkId
});
vaults[vaultId] = vault;
emit VaultBuilt(vaultId, owner, seriesId, ilkId);
}
Did you find it? Great, this was a real bug find in a bug bounty, so if you found it, it means you are able to already earn money through bounties, here is the explaination for those who didn't find it yet:
The vaultID for a new vault being built is required to be specified by the user building a vault via the build() function. An attacker can observe a build() as part of a batch transaction in the mempool, identify the vaultID being requested and front-run that by constructing a malicious batch transaction with only the build operation with that same vaultID. The protocol would create a vault with that vaultID and assign attacker as its owner. More importantly, the valid batch transaction in the mempool which was front-run will later fail to create its vault because that vaultID already exists, as per this check:
require (vaults[vaultId].seriesId == bytes6(0), "Vault already exists"); // Series can't take bytes6(0) as their id
As a result, the valid batch transaction fails entirely because of the attacker front-running with the observed vaultID.
While the attacker gains nothing except the ownership of an empty vault after spending the gas, this could grief the protocol’s real users by preventing them from opening a vault and interacting with the protocol in any manner.
Rationale for Medium-severity impact: While the likelihood of this may be low, the impact is high because valid vaults will never be successfully created and will lead to a DoS against the entire protocol’s functioning. So, with low likelihood and high impact, the severity (according to OWASP) is medium.
Example 3: Real life Cod4rena Bug Find With Payout (High Risk Vulnerability)
See if you can find the Front running vulnerability in the code bellow! This one is harder to spot. This code snippet is taken from a DEX (decentralized exchange)
// Calculate output of swapping SPARTA for TOKEN & update recorded amounts
function _swapBaseToToken(uint256 _x) internal returns (uint256 _y, uint256 _fee){
uint256 _X = baseAmount;
uint256 _Y = tokenAmount;
_y = iUTILS(_DAO().UTILS()).calcSwapOutput(_x, _X, _Y); // Calc TOKEN output
uint fee = iUTILS(_DAO().UTILS()).calcSwapFee(_x, _X, _Y); // Calc TOKEN fee
_fee = iUTILS(_DAO().UTILS()).calcSpotValueInBase(TOKEN, fee); // Convert TOKEN fee to SPARTA
_setPoolAmounts(_X + _x, _Y - _y); // Update recorded BASE and TOKEN amounts
_addPoolMetrics(_fee); // Add slip fee to the revenue metrics
return (_y, _fee);
}
The Function does not implement any slippage checks with comparing the swap / liquidity results with a minimum swap / liquidity value.
Users can be frontrun and receive a worse price than expected when they initially submitted the transaction. There's no protection at all, no minimum return amount or deadline for the trade transaction to be valid which means the trade can be delayed by miners or users congesting the network, as well as being sandwich attacked - ultimately leading to loss of user funds.
There are no minimum amounts out, or checks that front-running/slippage is sufficiently mitigated. This means that anyone with enough capital can force arbitrarily large slippage by sandwiching transactions, close to 100%.
Example 4: Personal Bug Bounty Find, Payout $2,500. Medium Risk
This is a bug I found while doing bug bounties, here is the code bellow, although easily finding thing the bug may need more context to the protocol, this code snipped should be enough. See if you can find the vulnerability:
function bondForWithHint(
uint256 _amount,
address _owner,
address _to,
address _oldDelegateNewPosPrev,
address _oldDelegateNewPosNext,
address _currDelegateNewPosPrev,
address _currDelegateNewPosNext
) public whenSystemNotPaused currentRoundInitialized {
// the `autoClaimEarnings` modifier has been replaced with its internal function as a `Stack too deep` error work-around
_autoClaimEarnings(_owner);
Delegator storage del = delegators[_owner];
uint256 currentRound = roundsManager().currentRound();
// Amount to delegate
uint256 delegationAmount = _amount;
// Current delegate
address currentDelegate = del.delegateAddress;
// Current bonded amount
uint256 currentBondedAmount = del.bondedAmount;
if (delegatorStatus(_owner) == DelegatorStatus.Unbonded) {
// New delegate
// Set start round
// Don't set start round if delegator is in pending state because the start round would not change
del.startRound = currentRound.add(1);
// Unbonded state = no existing delegate and no bonded stake
// Thus, delegation amount = provided amount
} else if (currentBondedAmount > 0 && currentDelegate != _to) {
// Prevents third-party caller to change the delegate of a delegator
require(msg.sender == _owner || msg.sender == l2Migrator(), "INVALID_CALLER");
// A registered transcoder cannot delegate its bonded stake toward another address
// because it can only be delegated toward itself
// In the future, if delegation towards another registered transcoder as an already
// registered transcoder becomes useful (i.e. for transitive delegation), this restriction
// could be removed
require(!isRegisteredTranscoder(_owner), "registered transcoders can't delegate towards other addresses");
// Changing delegate
// Set start round
del.startRound = currentRound.add(1);
// Update amount to delegate with previous delegation amount
delegationAmount = delegationAmount.add(currentBondedAmount);
decreaseTotalStake(currentDelegate, currentBondedAmount, _oldDelegateNewPosPrev, _oldDelegateNewPosNext);
}
{
Transcoder storage newDelegate = transcoders[_to];
EarningsPool.Data storage currPool = newDelegate.earningsPoolPerRound[currentRound];
if (currPool.cumulativeRewardFactor == 0) {
currPool.cumulativeRewardFactor = cumulativeFactorsPool(newDelegate, newDelegate.lastRewardRound)
.cumulativeRewardFactor;
}
if (currPool.cumulativeFeeFactor == 0) {
currPool.cumulativeFeeFactor = cumulativeFactorsPool(newDelegate, newDelegate.lastFeeRound)
.cumulativeFeeFactor;
}
}
// cannot delegate to someone without having bonded stake
require(delegationAmount > 0, "delegation amount must be greater than 0");
// Update delegate
del.delegateAddress = _to;
// Update bonded amount
del.bondedAmount = currentBondedAmount.add(_amount);
increaseTotalStake(_to, delegationAmount, _currDelegateNewPosPrev, _currDelegateNewPosNext);
if (_amount > 0) {
// Transfer the LPT to the Minter
livepeerToken().transferFrom(msg.sender, address(minter()), _amount);
}
emit Bond(_to, currentDelegate, _owner, _amount, del.bondedAmount);
}
Did you find it? Congratulations, this was classified as a medium risk finding with a payout of $2,500. If you found it then you are able to make money online.
Here is the bug description:
In the current implementation of the bondForWithHint function in the provided smart contract, there is a potential for a front-running attack that may force a user to unintentionally become a transcoder, preventing them from delegating their tokens to other users.
This can be achieved by an attacker front-running a transaction where a user (let's call this user 'Adam') is trying to delegate tokens to another user (let's say, 'Bob') using the bondWithHint (or bondForWithHint) function. The attacker can front-run Adam's transaction by calling bondForWithHint, setting _owner to Adam's address, _to to Adam's address again, and the bond amount to any non-zero value (even as low as 1 wei).
This would result in Adam unexpectedly becoming a transcoder because they now satisfy the check in the isRegisteredTranscoder function:
function isRegisteredTranscoder(address _transcoder) public view returns (bool) {
Delegator storage d = delegators[_transcoder];
return d.delegateAddress == _transcoder && d.bondedAmount > 0;
}
Because Adam is now a registered transcoder, their initial attempt to delegate to Bob using the bondWithHint function is blocked by the following check in the bondForWithHint function:
else if (currentBondedAmount > 0 && currentDelegate != _to) {
require(!isRegisteredTranscoder(_owner), "registered transcoders can't delegate towards other addresses");
The message "registered transcoders can't delegate towards other addresses" will be triggered, and the delegation will fail.
A user unknowingly becomes a transcoder and is unable to delegate their tokens to another user as initially intended. This may cause confusion for the user and disrupt the network's normal functioning, as the ability to freely delegate is a crucial aspect of the protocol. More importantly, it can be used by malicious actors to disrupt the delegation process on a larger scale if used repetitively.
Example 5: Personal Bug Bounty Find, Payout $20,000. High Risk
This is a bug I found while doing bug bounties, here is the code bellow, In the context of the protocol this was classified as High risk with a payout of $20,000, although the payout could have been $50,000, however the severity was downgraded due to the difficulty of the attack. See if you can find the issue
The context of this protocol is that every user that is on this protocol is essentially "staked in" so if fraudulent activity is detected, it is reported and if found guilty, the governing body can penalize the user by taking their tokens away from them.
struct Report {
address reporter;
address reportedAddress;
address secondReportedAddress;
uint256 reportTimestamps;
ILERC20 reportTokens;
bool secondReports;
bool reporterClaimStatus;
}
/// @notice This function will generate a report
/// @dev This function must be called by a non blacklisted/reported address.
/// It will generate a report for an address linked to a token.
/// Lossless Contracts, Admin addresses and Dexes cannot be reported.
/// @param _token Token address of the stolen funds
/// @param _account Potential malicious address
function report(ILERC20 _token, address _account) override public notBlacklisted whenNotPaused returns (uint256){
require(_account != address(0), "LSS: Cannot report zero address");
require(!losslessController.whitelist(_account), "LSS: Cannot report LSS protocol");
require(!losslessController.dexList(_account), "LSS: Cannot report Dex");
uint256 reportId = tokenReports[_token].reports[_account];
require(reportId == 0 ||
reportInfo[reportId].reportTimestamps + reportLifetime < block.timestamp ||
losslessGovernance.isReportSolved(reportId) &&
!losslessGovernance.reportResolution(reportId), "LSS: Report already exists");
reportCount += 1;
reportId = reportCount;
reportInfo[reportId].reporter = msg.sender;
tokenReports[_token].reports[_account] = reportId;
reportInfo[reportId].reportTimestamps = block.timestamp;
reportInfo[reportId].reportTokens = _token;
require(stakingToken.transferFrom(msg.sender, address(this), reportingAmount), "LSS: Reporting stake failed");
losslessController.addToBlacklist(_account);
reportInfo[reportId].reportedAddress = _account;
losslessController.activateEmergency(_token);
emit ReportSubmission(_token, _account, reportId, reportingAmount);
return reportId;
}
/// @notice This function will add a second address to a given report.
/// @dev This funtion must be called by a non blacklisted/reported address.
/// It will generate a second report linked to the first one created.
/// This can be used in the event that the malicious actor is able to frontrun the first report by swapping the tokens or transfering.
/// @param _reportId Report that was previously generated.
/// @param _account Potential malicious address
function secondReport(uint256 _reportId, address _account) override public whenNotPaused {
require(_account != address(0), "LSS: Cannot report zero address");
require(!losslessGovernance.isReportSolved(_reportId) && !losslessGovernance.reportResolution(_reportId), "LSS: Report already solved");
require(!losslessController.whitelist(_account), "LSS: Cannot report LSS protocol");
require(!losslessController.dexList(_account), "LSS: Cannot report Dex");
Report storage queriedReport = reportInfo[_reportId];
uint256 reportTimestamp = queriedReport.reportTimestamps;
ILERC20 token = queriedReport.reportTokens;
require(_reportId != 0 && reportTimestamp + reportLifetime > block.timestamp, "LSS: report does not exists");
require(queriedReport.secondReports == false, "LSS: Another already submitted");
require(msg.sender == queriedReport.reporter, "LSS: invalid reporter");
queriedReport.secondReports = true;
tokenReports[token].reports[_account] = _reportId;
losslessController.addToBlacklist(_account);
queriedReport.secondReportedAddress = _account;
emit SecondReportSubmission(token, _account, _reportId);
}
/// @notice This function solves a report based on the voting resolution of the three pilars
/// @dev Only can be run by the three pilars.
/// When the report gets resolved, if it's resolved negatively, the reported address gets removed from the blacklist
/// If the report is solved positively, the funds of the reported account get retrieved in order to be distributed among stakers and the reporter.
/// @param _reportId Report to be resolved
function resolveReport(uint256 _reportId) override public whenNotPaused {
require(!isReportSolved(_reportId), "LSS: Report already solved");
(,,,uint256 reportTimestamps,,,) = losslessReporting.getReportInfo(_reportId);
if (reportTimestamps + losslessReporting.reportLifetime() > block.timestamp) {
_resolveActive(_reportId);
} else {
_resolveExpired(_reportId);
}
reportVotes[_reportId].resolved = true;
delete reportedAddresses;
emit ReportResolve(_reportId, reportVotes[_reportId].resolution);
}
/// @notice This function is for the reporter to claim their rewards
/// @param _reportId Staked report
function reporterClaim(uint256 _reportId) override public whenNotPaused {
require(reportInfo[_reportId].reporter == msg.sender, "LSS: Only reporter");
require(losslessGovernance.reportResolution(_reportId), "LSS: Report solved negatively");
Report storage queriedReport = reportInfo[_reportId];
require(!queriedReport.reporterClaimStatus, "LSS: You already claimed");
queriedReport.reporterClaimStatus = true;
uint256 amountToClaim = reporterClaimableAmount(_reportId);
require(queriedReport.reportTokens.transfer(msg.sender, amountToClaim), "LSS: Token transfer failed");
require(stakingToken.transfer(msg.sender, reportingAmount), "LSS: Reporting stake failed");
emit ReporterClaim(msg.sender, _reportId, amountToClaim);
}
Did you find it? Great, that means you have the skills required to earn $20,000 - $50,000 on a single bug report! For those who didnt find the bug, don't worry, here is a description:
Once a user has called the function report, the user can then call secondReport and input a different address to be penalized and subsequently have its tokens removed. This presents a very nasty front-running risk, but its not the reporter being front-run....its the reporter doing the front-running! This is because after a report is voted on by the governance and validated, when the governance calls resolveReport, the reporter can listen to this contract call, and front run this transaction by calling secondReport and inputting the address of a victim (e.g. a whale or a dex). Because this transaction is completed before 'resolveRepot', 'resloveReport' now pulls the tokens from the victim and distributes it to everybody involved, the proposedWallet, stakers, community members.
