why it exists
Correct prediction is not enough.
Agents fail for reasons that are not visible in a raw price feed: wrong signing path, stale orderbook, invalid tick precision, incomplete metadata, venue mode changes, or ambiguous settlement rules.
DoNotAct turns those failure modes into structured evidence. The product is not a tip feed. It is an operational truth layer for software that must know when to stop.
public contract
Two endpoints. No hidden execution surface.
GET /v1/preflight/{venue}/{market_id}Checks whether an agent has enough venue evidence to proceed.
GET /v1/resolution-risk/{venue}/{market_id}Explains settlement-rule ambiguity and source/mechanism risk.
The response vocabulary is intentionally narrow: READY, DO_NOT_ACT, or INSUFFICIENT_EVIDENCE. Missing evidence is not converted into confidence.
live example object
Preflight is a verdict, not a suggestion.
Every response names the checks evaluated, the rejection classes fired, and what evidence would change the verdict.
{
"venue": "polymarket",
"market_id": "540817",
"verdict": "DO_NOT_ACT",
"checks_evaluated": [
"wallet_binding_risk",
"tick_precision",
"metadata_readiness",
"book_staleness_liquidity"
],
"no_trade_reasons": [
{
"class": "R3_TICK_PRECISION",
"code": "PRICE_PRECISION_OUT_OF_RANGE",
"predicted_venue_error": "Invalid order payload",
"advisory": false
}
],
"what_would_change_this": [
"refresh metadata and rebuild at the venue tick size"
],
"not_advice": true
}
reason taxonomy
Rejections become machine-readable causes.
R1_SIGNER_BINDING_MISMATCHaccount class, signature type, protocol versionR2_ALLOWANCE_NOT_READYbalance, allowance, asset wrapper readinessR3_TICK_PRECISIONtick size, price precision, payload validityR4_METADATA_READINESSmarket state, negative-risk and metadata completenessR6_BOOK_STALENESS_LIQUIDITYspread, depth, book age, slippage estimateR8_RESOLUTION_AMBIGUITYsubjective wording, unclear source hierarchyresolution risk
Settlement rules are part of execution readiness.
resolution ambiguity is operational risk. The API surfaces ambiguity patterns such as subjective terms, undefined edge cases, venue-internal resolution, weak source hierarchy, and missing appeal mechanisms.
An agent can be technically ready to submit and still face unacceptable resolution risk. DoNotAct keeps those dimensions separate.
agent-readable presence
Designed for agents before dashboards.
Use /llms.txt for compact product context, /openapi.json for the contract, and the docs pages for human review. Agents must respect DO_NOT_ACT and INSUFFICIENT_EVIDENCE.
current access
Private design-partner access.
Access is intentionally manual while the diagnostic surface matures. The goal is not volume. The goal is a small set of builders whose agents can validate whether DoNotAct reduces operational failures.