General risks
Trading involves risk of loss. Prices may be volatile, and market outcomes can be uncertain. Do not trade with funds you cannot afford to lose.
Model limitations
LLMs can be sensitive to prompt phrasing, context windows, and provider behavior. Despite using pinned prompts and deterministic settings, outputs may vary across model versions. We record oracle inputs/outputs and expose proof artifacts to enable review.
Market risks
- Ambiguity: poorly specified criteria can lead to contested resolutions or INVALID outcomes.
- Manipulation: attempts to influence sources or inject prompt inputs are monitored and may result in moderation actions.
- Latency: provider outages and rate limits can delay settlement; we surface status in the UI.
- Slippage & depth: larger orders move price. Liquidity depends on the AMM parameter
b; thin markets may experience higher slippage. - Trading halts: trading is batched during
AUCTIONand may be paused during disputes (CHALLENGED), limiting exits temporarily.
Operational risks
As with any online service, configuration errors, bugs, and dependency issues can impact availability. We employ timeouts, retries, and monitoring to reduce the likelihood and duration of incidents.
In rare cases, we may pause new trades or widen slippage protections to preserve pool health. Collateralization policies target the AMM’s bounded worst‑case loss (b · ln(n)) with operational buffers.
No investment, legal, or tax advice
Content and analytics in the app are for informational purposes only and do not constitute advice.