DeXposure-Claw: An Agentic System for DeFi Risk Supervision
For financial regulators supervising decentralized finance, this system addresses the critical problem of high false-intervention rates in LLM-based risk monitoring, offering a more reliable and auditable approach.
DeXposure-Claw is a forecast-grounded agentic system for DeFi risk supervision that reduces false alarms by routing LLM decisions through structured evidence from a graph time-series foundation model, deterministic monitors, and confidence gates. Experiments on five years of weekly real data show it outperforms general-purpose LLM agents in regulator-aligned evaluations.
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. 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. 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 https://github.com/EVIEHub/DeXposure-Claw.