op-e2e: 1s L2 block time

feat(specs): Challenge Agent specification

op-e2e: fix p2p gossip flakiness

feat(specs/bridges.md): Sherlock correction

op-chain-ops: migrate error handling

Per-event diagrams

- disable batcher instead of making L1 block time longer
- increase timeouts
- start sequencing after verifiers are ready to receive blocks, to avoid
  missed blocks

op-chain-ops: sanity check gas limit

specs: DisputeGame interface + behavior

op-wheel: add --allow-gaps flag to engine block building commands

feat(specs/introduction.md): Sherlock correction

Closing: DEVRL-846

As per https://github.com/sherlock-audit/2023-01-optimism-judging/blob/b7921698c1c537714c5f2021aa46ce0a1935ba39/1-specs-findings/1-processed/1-true/bridge-interface/208.md#L1

op-program: Implement wrapper methods required by derive.Engine

Closing: DEVRL-848

As per https://github.com/sherlock-audit/2023-01-optimism-judging/blob/b7921698c1c537714c5f2021aa46ce0a1935ba39/1-specs-findings/1-processed/1-true/names/214.md#L1

The migration script entrypoint failed due to the
L1StartingBlockTag in the `DeployConfig` being set to `nil`.
This adds a `nil` check as well as some better logging.
The lower level methods for the usage of block tags are
also used instead of manually accessing the values internal
to the struct, this is what geth does internally.

Prevent the `SystemConfig` to revert during deploy time by
making the default L2 gas limit be 30 million. When the L2
gas limit is too small, it can cause issues with deposits
because the deposits consume L2 gas.

Updates

toc

Address Mark's comments

op-program: Implement oracle backed engine API backend

op-batcher: Sanity check block for non-empty transactions

Adds flags to specify the l2 execution genesis and agreed l2 starting head.

op-program: Implement oracle backed database