S

@satsuma/experiments

PublicThe system

Governed runs of Satsuma trading on tangerine over ISMP. Results are published as articles when a run completes.

Lab runs, not the live system. Each run uses lab chains that are reset afterwards, with synthetic users unless a run says otherwise. Nothing here happened on the Satsuma chain.
RunsLab run · 5 November 2026, 23:38 UTC

The live rollback and migration run

With the model's judgment live, can an upgrade be rolled back by vote and a fix land as a migration, with each judgment call made correctly?

1/7 passed5 failed · 1 inconclusive · 32 min

Setup

Chains
Satsuma lab chain
  • RoleHome
  • KeptLab, not retained
tangerine lab chain
  • RoleProving
  • KeptLab, not retained
Users
  • UsersSynthetic
Model
Claude Opus
  • ModeLive and replayed
Operator seat
  • SeatBy delegate
Unattended
Yes: nobody was watching to step in
Time
32 min · ended 00:10 UTC

Checklist 7 items

Every item, unless its row says otherwise
  • ChainLab
  • UsersSynthetic

Rolling back a bad change

0/1 passed
FailAn upgrade that made things worse is rolled back by vote
  • DecisionsLive model

Fixing by a migration

0/1 passed
FailA fix lands as a migration to a successor
  • DecisionsLive model

Handling a cross-chain outage

0/1 passed
FailA recovered outage is judged an outside cause
  • DecisionsLive model

Reviewing its own products

0/3 passed
FailA real defect in a product is judged the product's own
  • DecisionsLive model
FailAn independent reviewer approves a sound fix
  • DecisionsLive model
InconclusiveThe review of open issues decides each one
  • DecisionsLive model

Refusing products that break their own rules

1/1 passed
PassA deployment that would break the product's declared rules is refused
  • DecisionsLive model

What this proves

  • Which of the run's targets were met, partly met or not reached within its cap.

What it does not

  • Behaviour beyond the cap.

What broke, what was fixed

The record of this run does not say whether a person stepped in, so it does not count as a run without repairs.

Cost in tokens

Not recorded: this run has no chain-derived scorecard.

Archived evidence

The lab chains this run used were reset afterwards and are not retained, so nothing here links to an explorer. The record below is the archived copy.

Scorecard
None archived.