Choose the job

Three categories share one search term.

Market platforms create contracts, manage liquidity, and settle outcomes. Analytics tools compare probabilities, positions, or activity across venues.

Trading agents consume market data and produce an order or explicit refusal. They need stronger controls because a sound forecast can still become a poor trade.

Six checks before you automate

CheckWhat good software showsWarning sign
Market dataNamed venues, current prices, close times, and outcome rulesProbabilities without source or timestamp
ExecutionExecutable quotes, fees, slippage, and price impact before submissionBacktests using midpoint prices as fills
EvidenceDeclared sources, freshness limits, and exact settlement mappingFree-form claims that cannot be replayed
RiskSpend, bankroll, size, impact, and stale-quote limitsOne confidence score controls the order
RecoveryDuplicate protection and explicit ambiguous-order handlingBlind retries after a lost response
ProofDecision receipts through settlement and realized resultsSimulated profit presented as live performance

Automation boundary

No trade is a valid output.

Reliable software fails closed when evidence, quotes, or transaction state is stale or unclear. It does not turn missing data into certainty.

Ask how the system handles restarts, partial responses, settlement, and redemption. These paths decide whether performance claims can be trusted.

Where Settlement Edge fits

A narrow trading agent, not a market platform.

Settlement Edge is open-source TypeScript software for the closed Gensyn Delphi Agent Arena competition. It validates reviewed facts, requests quotes, applies strict caps, and hash-links decisions.

It is not a hosted prediction market, analytics dashboard, mainnet system, or proven profit engine. The final record shows 0 submitted orders and 0.0000 TST realized P&L.

Evidence and next read

Product scope and observed result checked August 26, 2026.