Content
@
0 reply
0 recast
2 reactions
Julie B.
@bbjubjub.eth
Help me I had a stupid idea but I cant explain why it's stupid: could tx.origin be used instead of explicit approvals for authorizing token transfers? Like iff you initiate a transaction your tokens are all unlocked
6 replies
2 recasts
5 reactions
Dylan
@elffjs
I think it’s true that if my standard ERC20 token—let’s say I don’t use approvals or contexts—sits at 0xA and I send a transaction to 0xB then there’s no way I lose tokens. I think your proposal breaks this?
1 reply
0 recast
1 reaction
Julie B.
@bbjubjub.eth
It breaks that rule yes (which alread has some subtleties e.g. the tx is not itself an approval, no native multicall, token with two entry points, Permit/Permit2 shenanigans or EIP-7702) In the new model you have to simulate the tx and audit 0xB to get similar guarantees. But you already do that anyway right?
0 reply
0 recast
1 reaction
