JevLoop

routes the loop's own judgements to Jev, where a `Choice` picks the next tool from candidates rebuilt every step, a `Score` grades the call's risk, and a `Noul` decides whether it needs authorisation, while plain code acts on the answers so a high risk score forces human authorisation that no probability can override (7.7% of wall clock with the offline judge, 79% over the hosted API).

CHOICE 调用长什么样

这是该提问类型的通用骨架，不是这个项目的真实代码。

from typesafe import TypeSafe

ts = TypeSafe()  # reads TYPESAFE_API_KEY
result = ts.evaluate(
    state=page_state,
    questions={
        "next_action": {
            "type": "choice",
            "options": ["click", "scroll", "type", "done"],
            "instructions": "What should the agent do next?",
        }
    },
)
action = result["next_action"]          # the chosen option
confidence = result["next_action_confidence"]