Auditor detailed a front-running vulnerability in FairWin

Researchers detailed a FairWin front-running flaw

On September 29, 2019, smart-contract auditor and Kleros chief technology officer Clément Lesaege publicly described a front-running vulnerability in FairWin, an Ethereum application that accepted ether deposits and distributed referral rewards and advertised dividends. The disclosure said an attacker could watch a pending deposit, copy its invitation code and submit a competing transaction with a higher gas price so miners would execute the attacker’s transaction first.

That ordering mattered because, according to Lesaege’s review of the contract, an invitation code became associated with the first address that used it. A later deposit carrying the same code could then be credited to that first address. The mechanism turned Ethereum’s public transaction queue—the mempool—from a transparency feature into an attack surface: visibility before confirmation gave an observer an opportunity to reorder who received attribution.

The September 29 post included links to the deployed contract and extracted Solidity code. It was a technical disclosure by a named auditor, not a finding by a regulator or an independent incident-response firm. It established a plausible code path and warned that funds were at risk; it did not establish that anyone had successfully used that path.

The exposure extended beyond one bug

Lesaege’s analysis identified a second, more centralized risk. FairWin’s operator controlled execution of rewards, dividends and awards. The auditor argued that the operator could favor selected accounts, withhold other payouts and progressively drain the contract through accounts it controlled. Users could also be prevented from reaching the sequence of rewards needed to recover their deposits.

The front-running path itself came with constraints. The attacker’s transaction had to land before the intended depositor’s transaction, the operator had to continue processing payouts, and the contract had to remain solvent through the waiting period. Lesaege stated that users could withdraw five days after depositing, subject to sufficient funds remaining. Those conditions reduced the certainty of profit for an attacker but did not repair the attribution error or the operator’s discretion.

FairWin’s response, reproduced by Lesaege as an update to the disclosure, denied that the behavior was a vulnerability and said the project monitored for intruders. That was an unattributed project response relayed by the researcher, not a verified remediation. No patched deployment, formal audit or on-chain proof of a fix accompanied it on September 29.

Why the disclosure mattered

The episode illustrated two distinct forms of smart-contract risk. Front-running arose from transaction ordering and contract logic. Administrative risk arose because a nominally autonomous application still gave its operator decisive control over distributions. Code on a public chain made both structures inspectable, but publication alone could not stop pending transactions, rewrite the contract or guarantee users could withdraw.

The disclosure also showed why “decentralized” was an incomplete description of an application. Ethereum settled calls to FairWin, but settlement did not validate its economic design, remove privileged functions or ensure that referral accounting was safe. On September 29, the defensible conclusion was narrower: researchers had documented serious vulnerabilities in a heavily used contract, while exploitation, operator intent and ultimate losses remained unresolved.

Later context

On October 1, researcher Philippe Castonguay published a retrospective account saying the contract held about 19,000 ETH when the detailed September 29 disclosure appeared and zero ETH on September 30. He reported no evidence that the front-running flaw had been exploited and could not rule out preferential operator behavior. Those balances and conclusions are later context, not facts available with the same certainty during the September 29 disclosure.