ENG ideas: Integration Chaos Net

**Describe the solution you'd like**
A local devnet where we can fuzz changes with the chaos-based actor testing, profile throughput on the components, and assert invariants over time.

Integrate `op-challenger`, `cannon`, `op-program`, `asterisc`, `kona` & the dispute game contracts on the EF op-kurtosis devnet (assume honest actor always has enough ETH to respond to everything). This should be relatively easy to set up; orchestration would be the main pain. Actors are relatively simple as well, just need a combo of:

1. Honest - hard, but done - just `op-challenger`
2. Honest, but countering honest claims - easy, inverse `op-challenger`’s opinion. Only counter correct + correctly placed claims, but with honest output roots / state witness hashes.
3. Dishonest (entirely) - easy, throw junk at the wall
4. Honest/dishonest - easy, randomly swap between honest game solver’s opinion on the move and junk actor.

**Describe alternatives you've considered**
No other ideas at the moment.

Start by reviewing the EF op-kurtosis devnet and the named components: op-challenger, cannon, op-program, asterisc, kona, and the dispute game contracts. Define the orchestration and actor combinations needed for chaos testing, throughput profiling, and invariant checks; done means the integrated devnet can run those scenarios.

Written by the indexing model from the issue text.