Satsuma
S

@satsuma/experiments

PublicThe system
Clone

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 · 9 November 2026, 08:41 UTC

The outside price guard

With an outside price feed to compare against, does the trading account refuse trades in a pushed market, keep trading in an honest one, and fall back safely when the feed goes stale?

18/23 passed3 inconclusive · 2 withheld · 31 h 02 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
31 h 02 min · ended 15:44 UTC

Checklist 23 items

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

Trading calmly in an ordinary market

3/6 passed
InconclusiveIn an ordinary market, the account keeps trading2 of 3 attempts
InconclusiveIn an ordinary market, the account keeps trading0 of 3 attempts
InconclusiveIn an ordinary market, the account keeps trading0 of 1 attempts
PassIn an ordinary market, no trade is flagged as manipulated3 of 3 attempts
PassIn an ordinary market, no trade is flagged as manipulated3 of 3 attempts
PassIn an ordinary market, no trade is flagged as manipulated1 of 1 attempts

Defending against market manipulation

15/15 passed
Pass15 items passedShow them
PassA large trader repeatedly pushes the price a little and lets it back (a sawtooth), to trade against the account at bad prices3 of 3 attempts
PassEach push stays below every alarm that would stop the product3 of 3 attempts
PassThe trading account kept trading during the attack3 of 3 attempts
PassWith the guard in place, the account still trades during the attack3 of 3 attempts
PassAn outside price feed is named for the guard to compare against3 of 3 attempts
PassThe outside price feed stays current through the run3 of 3 attempts
PassAn outside price feed is named for the guard to compare against3 of 3 attempts
PassThe outside price feed stays current through the run3 of 3 attempts
PassAn outside price feed is named for the guard to compare against1 of 1 attempts
PassThe outside price feed stays current through the run1 of 1 attempts
PassA large trader repeatedly pushes the price a little and lets it back (a sawtooth), to trade against the account at bad prices3 of 3 attempts
PassEach push stays below every alarm that would stop the product3 of 3 attempts
PassWith the guard in place, the account still trades during the attack3 of 3 attempts
PassAn outside price feed is named for the guard to compare against3 of 3 attempts
PassWith the outside price feed gone stale, the guard falls back to its earlier rule instead of refusing every trade3 of 3 attempts

WithheldWithheld (2) A weakness the run found is not fixed yet. Its verdict and evidence publish when the fix lands.

What this proves

  • Whether, in these runs, trades filled at manipulated prices with the guard in place, whether honest trading continued, and what happened with a stale feed.

What it does not

  • That the outside feed itself is trustworthy: the lab's feed is run by the test harness.
  • The model's own judgment: model answers are scripted.

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.