Hi @frangio,
I’ve been working extensively on accounts, and I remain convinced that sophisticated account abstraction use cases will require more robust schemes, such as ERC-7803, rather than ERC-7739
Particularly, cross-chain account interoperability seems more approachable if there are ways to rely on EIP-712 domains to construct cross-chain operations (e.g., intents). One strategy could be using chainId(0)
to create intentionally replayable signatures across chains. I think these problems are more tractable with ERC-7803’s signing domain approach. See Universal Cross-Chain Signatures for Account Abstraction
On the security considerations side, I haven’t identified any particular concerns with the core mechanism. However, if this standard becomes a foundation for cross-chain setups, there may be some security aspects worth exploring (e.g., is using signing domains with chainId(0)
the best approach for multi-chain signatures?).
I opened a PR expanding on the rationale and backwards compatibility sections. Would love to hear your thoughts.
Another question, what do you think of enforcing ERC-{n}
in the id
field so that so that ECDSA
is covered by ERC-7913?