Submitted by gpersoon, also found by leastwood and ScopeLift
The function bondToAccount() of Bonding.sol has a check based on _notSameBlock()
_notSameBlock() makes sure the same msg.sender cannot do 2 actions within the same block.
However this can be circumvented in this case:
Suppose you call bondToAccount() via a (custom) smart contract, then the msg.sender will be the address of the smart contract.
For a pseudo code proof of concept see below.
Im not sure what the deeper reason is for the _notSameBlock() in bondToAccount().
But if it is important then circumventing this check it will pose a risk.
call function attack1.attack()
https://github.com/code-423n4/2021-11-malt/blob/d3f6a57ba6694b47389b16d9d0a36a956c5e6a94/src/contracts/Bonding.sol#L81-L92
https://github.com/code-423n4/2021-11-malt/blob/d3f6a57ba6694b47389b16d9d0a36a956c5e6a94/src/contracts/Permissions.sol#L135-L141
Add access controls to the function bondToAccount()
An end-user could still call bond()
0xScotch (sponsor) confirmed 
Alex the Entreprenerd (judge) commented:
