You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Upon the market creation, ERC20 tokens are deployed respresenting an exact outcome. A bytecode tied to a certain outcome depends on tokenID which itself depends on:
collateralToken
collectionID
This means that there's a possibility to front-run the market creation by changing only openingTime, which results in having the same tokenID for each outcome for both markets within Seer, but on reality, we'll have 2 different questions with openingTime being deviated from each other.
Attack Scenario
Someone creates a categorial market X representing a chess match with a valid openingTime(in 2 hours should be resolved)
An adversary sees it, and creates an identical market Y on Seer, which itself creates another question on reality.eth, but with openingTime specified as delta < X.openingTime.
Since the erc20 tokens deployed for outcomes are the same, an adversary can finalize his market with a desired outcome receiving his collateral back and sells other outcomes on amm.
As a result, he's able to extract free-risk value.
Mitigation Steps
There should not be cases, where two different question_ids on reality.eth having the same erc20 positions deployed.
The text was updated successfully, but these errors were encountered:
Github username: @Rassska
Twitter username: m_Rassska
Submission hash (on-chain): 0x2af608995e56e59bedc52897b80e6b7a646f814436b166de2c8562e1c02cc465
Severity: high
Description:
Description
Upon the market creation, ERC20 tokens are deployed respresenting an exact outcome. A bytecode tied to a certain outcome depends on tokenID which itself depends on:
collateralToken
collectionID
This means that there's a possibility to front-run the market creation by changing only openingTime, which results in having the same tokenID for each outcome for both markets within Seer, but on reality, we'll have 2 different questions with openingTime being deviated from each other.
Attack Scenario
Mitigation Steps
The text was updated successfully, but these errors were encountered: