We do slashing checks on sync committees

[GalRogozinski]

As part of committee consensus we perform attestation slashing checks on combined attestation/sync-committee data.

Meaning that slashable attestation data will hinder sync committee duty.

Since slashable data is rare and this should optimistically be resolved on the next QBFT round, it isn't this bad.

Originally posted by @GalRogozinski in #421 (comment)

[GalRogozinski]

This needs to be discussed

[GalRogozinski]

So we have a max of 512 sync committee participants. As the network grows this number will become more and more negligible.

So perhaps it makes sense to separate runners for attestations and sync committees so we have a clean protocol that won't run into rare issues.

On the other hand, we can have an optimization instead that all non-slashable duties will use the same consensus

[GalRogozinski]

This is easily solvable
In spec we need to initialize valueCheckF with only the validators that do attestation duties

[GalRogozinski]

Solved by #471