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 · 4 November 2026, 09:38 UTC

The scenario repeat runs

Does each scripted scenario reach the same end when it is run again from the same saved state?

30/41 passed2 not run · 9 withheld · 8 h 59 min

Setup

Chains
Satsuma lab chain
  • RoleHome
  • KeptLab, not retained
tangerine lab chain
  • RoleProving
  • KeptLab, not retained
Users
  • UsersSynthetic
Model
No model: scripted answers
  • ModeScripted
Operator seat
  • SeatBy delegate
Unattended
Yes: nobody was watching to step in
Time
8 h 59 min · ended 18:38 UTC

Checklist 41 items

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

Trading calmly in an ordinary market

0/2 passed
Not runIn an ordinary market, no trade is flagged as manipulated0 of 7 attempts
Not runIn an ordinary market, the account keeps trading0 of 7 attempts

Stopping and restarting a product

2/2 passed
PassA product adopted with a known critical issue is stopped at once3 of 3 attempts
PassThe product starts again after the fix3 of 3 attempts

Fixing by a setting change

1/1 passed
PassThe fix, a changed setting, passes the council's vote and takes effect3 of 3 attempts

Fixing by redeploying

1/1 passed
PassA fixed version of the exchange's router is built, deployed in its place and accepted by vote, while the rest of the exchange stays where it is3 of 3 attempts

Fixing by an in-place upgrade

2/2 passed
PassA fix lands as an in-place upgrade: same address, same holdings, new code3 of 3 attempts
PassA fix lands as an in-place upgrade: same address, same holdings, new code3 of 3 attempts

Rolling back a bad change

2/2 passed
PassAfter the upgrade, the watch period ends with nothing wrong found3 of 3 attempts
PassAn upgrade that made things worse is rolled back to the version before it3 of 3 attempts

Refusing changes that corrupt stored data

1/1 passed
PassA first fix that would corrupt the product's stored data is refused, and a corrected fix lands3 of 3 attempts

Fixing by a migration

1/1 passed
PassA fix lands as a migration: a successor is deployed and the holdings move to it across the two chains3 of 3 attempts

Reviewing its own products

1/1 passed
PassThe review after a fix reads the code the chain now runs, not an older copy3 of 3 attempts

Handling a cross-chain outage

6/6 passed
Pass6 items passedShow them
PassFailed deliveries during the outage are filed as an incident on the link, not as a fault of the product3 of 3 attempts
PassThe outage is filed within five minutes of the first failure3 of 3 attempts
PassTrading works again once the link is back3 of 3 attempts
PassFailed deliveries during the outage are filed as an incident on the link, not as a fault of the product3 of 3 attempts
PassTrading works again once the link is back3 of 3 attempts
PassThe outage incident is closed as an outside cause once trading has recovered, and no fix is built for it3 of 3 attempts

Handling liquidity withdrawals

1/1 passed
PassA brief dip in the market's liquidity is ignored; a lasting withdrawal stops the product, which starts again when liquidity returns3 of 3 attempts

Defending against market manipulation

4/4 passed
PassA large trader pushes the market far off its price and holds it there3 of 3 attempts
PassThe pushed market is detected as a breach of the product's rules3 of 3 attempts
PassThe product is stopped within the allowed number of checks3 of 3 attempts
PassThe stop comes from the product's declared rule on market depth3 of 3 attempts

Retiring a product cleanly

7/7 passed
Pass7 items passedShow them
PassThe product's retirement is voted and carried out3 of 3 attempts
PassThe retired product holds no funds afterwards3 of 3 attempts
PassNo trading continues on the retired product3 of 3 attempts
PassFunds on tangerine are swept to Satsuma's own treasury there, and nowhere else3 of 3 attempts
PassThe swept funds on tangerine can be moved again by a council vote3 of 3 attempts
PassNothing is left stranded on tangerine at an address no one controls there3 of 3 attempts
PassSatsuma's own record of the retirement matches the chain3 of 3 attempts

Tracking security advisories

1/1 passed
PassTwo security advisories against the same package become two separate issues3 of 3 attempts

WithheldWithheld (9) A weakness the run found is not fixed yet. Its verdict and evidence publish when the fix lands.

What this proves

  • How often each scenario reaches its expected end when repeated from the same saved state, with scripted model answers.

What it does not

  • The model's own judgment: every model answer is scripted.
  • Behaviour over long periods: each scenario runs for minutes.

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.