The ACDT #73 discussion raised a question for me: if the committee is selected based on a previous epoch's state but validated against current slot balances, doesn’t the proposed Engine API support for older-head reorgs introduce a specific risk of consensus splits via state-dependent validator set drift? Curious if others see this timing game as a viable vector before the PTC lookahead bug is patched. @ethereum @EFDevcon

@potuz_eth