DeXposure-Claw: An Agentic System for DeFi Risk Supervision

DeXposure-Claw：用于 DeFi 风险监管的智能体系统

Abstract: Decentralized finance exposes supervisors to fast-moving, networked credit risks. General-purpose LLM agents fit this setting poorly: they over-read weak evidence and recommend high-stakes interventions, while existing evaluations offer no regulator-aligned way to measure the resulting false alarms.

摘要： 去中心化金融（DeFi）使监管者面临快速变化且相互关联的信用风险。通用大语言模型（LLM）智能体在这一场景下的表现并不理想：它们往往会过度解读微弱的证据并建议采取高风险的干预措施，而现有的评估方法无法提供一种符合监管要求的手段来衡量由此产生的误报。

We introduce DeXposure-Claw, a forecast-grounded agentic supervision system that routes LLM decisions through structured evidence: (1) DeXposure-FM, a graph time-series foundation model, forecasts future exposure networks; (2) deterministic monitors and stress scenarios then turn those forecasts into typed alerts, attribution signals, and scenario evidence; and (3) data-health and confidence gates constrain escalation before DeXposure-Claw emits auditable supervisory tickets with rationales.

我们引入了 DeXposure-Claw，这是一个基于预测的智能体监管系统，通过结构化证据引导 LLM 的决策过程：(1) DeXposure-FM（一种图时间序列基础模型）用于预测未来的风险敞口网络；(2) 确定性监控器和压力测试场景将这些预测转化为分类警报、归因信号和场景证据；(3) 数据健康检查和置信度门控机制在 DeXposure-Claw 发布带有逻辑说明的可审计监管工单之前，对升级行为进行约束。

We further develop DeXposure-Bench, a six-axis evaluation harness, whose decision axis scores tickets against a regulator-aligned absolute-loss ground truth and an explicit false-intervention rate. Experiments on five years of weekly real data fully support our system. Code is at this https URL.

我们进一步开发了 DeXposure-Bench，这是一个六轴评估框架，其决策轴根据符合监管要求的绝对损失基准和明确的错误干预率对工单进行评分。基于五年每周真实数据的实验充分验证了我们系统的有效性。代码链接见此处。