GalloDaSballo - tx.origin breaks ability to have someone else broadcast TXs and may cause loss of bonds

[sherlock-admin3]

GalloDaSballo

medium

tx.origin breaks ability to have someone else broadcast TXs and may cause loss of bonds

Summary

The use of tx.origin adds an additional risk to the bonding system that can be avoided by passing the payer in initialize

Vulnerability Detail

Impact

The FaultDisputeGame.initialize sets the starting claimData as follows:

https://github.com/sherlock-audit/2024-02-optimism-2024/blob/main/optimism/packages/contracts-bedrock/src/dispute/FaultDisputeGame.sol#L555-L565

        claimData.push(
            ClaimData({
                parentIndex: type(uint32).max,
                counteredBy: address(0),
                claimant: tx.origin,
                bond: uint128(msg.value),
                claim: rootClaim(),
                position: ROOT_POSITION,
                clock: LibClock.wrap(Duration.wrap(0), Timestamp.wrap(uint64(block.timestamp)))
            })
        );

Setting the claimant to tx.origin this can be used to:

• Steal a benign claimant bond by "sponsoring" their tx
• Makes the system incompatible with Account Abstraction
• Creates issues for Keepers (as the fee will be attributed to them instead of the actual payer)

The rest of the system doesn't prevent SCs or SC Wallets from interacting with it, but this setting does break composability

Code Snippet

• User A wants to get their withdrawals proven automatically by a 3rd party service
• The third party service will have their relayers EOAs steal funds and force them to work via a trusted system
• The third party EOA will be tx.origin, user a will pay the bond and not be guaranteed to receive it back

https://github.com/sherlock-audit/2024-02-optimism-2024/blob/main/optimism/packages/contracts-bedrock/src/dispute/FaultDisputeGame.sol#L559-L560

Tool used

Manual Review

Recommendation

A simple refactoring that forwards the original payer or specifying a bondReceiver would be sufficient in this case

Duplicate of #194

[sherlock-admin2]

The protocol team fixed this issue in the following PRs/commits:
ethereum-optimism/optimism#10149