BoldToken initialization can be frontrun to silently mint/approve bold to an attacker

Summary

BoldToken::initialize() is frontrunnable.

Vulnerability Detail

BoldToken::initialize() is called seperately after the constructor, which means that if sent in different txs (depending on deployment scripts, it may be possible to mitigate if sent via a multicall), an attacker can frontrun the deployment, initialize the BoldToken to a malicious implementation, mint/approve Bold to their address and then set the implementation to address(0) so the script does not revert and the BoldToken::initialize() call succeeds.

Impact

Attackers mints/gets Bold for free.

Code Snippet

https://github.com/sherlock-audit/2025-02-nerite/blob/main/nerite/contracts/src/BoldToken.sol#L59-L72 https://github.com/superfluid-finance/protocol-monorepo/blob/1edc4ed3ddafa87d42c1155e870532e2a5a80470/packages/ethereum-contracts/contracts/upgradability/UUPSProxy.sol#L25-L29

Tool Used

Manual Review

Recommendation

Use a multicall to deploy or move the BoldToken::initialize() logic to the constructor.