Klyr
Open Console
SECURITY

Klyr does not treat model output as authorization.

01

Model output is not authorization

An AI model may propose a structured claim, rank candidates, or extract resolution conditions. It never declares a market verified or approves execution.

02

Deterministic policy enforcement

Risk policy - position limits, liquidity thresholds, duplicate protection - is evaluated by fixed, auditable logic, not model judgment.

03

Fail-closed behavior

When market data is unavailable, resolution rules are ambiguous, or a required check cannot complete, Klyr responds with BLOCK or ESCALATE - never a silent guess.

04

Explicit ambiguity states

A vague deadline or unclear threshold is surfaced as an ambiguity, not resolved by assumption.

05

Execution idempotency

Every execution requires an idempotency key. A replayed request returns the original decision receipt rather than executing twice.

06

Bounded requests and timeouts

External calls are time-bounded. A timeout produces a typed, graceful failure rather than a hang.

07

Verification snapshots and receipts

Every decision preserves what Klyr actually knew at the time: the claim, the market, the resolution rules, the state, the policy result, and a decision hash.