Autonomous DeFi circuit breaker that auto-evacuates funds before pools collapse

Umbra: The shadow that guards your funds. Umbra is an autonomous financial circuit breaker designed to protect DeFi liquidity providers and traders from systemic market crashes and depegs around the clock. Because crypto markets never sleep, but users do, Umbra continuously monitors live pool data via The Graph to track sudden liquidity drains, volume spikes, and pricing anomalies. When a crisis begins, our quantitative risk model (the CBRI index, hosted and verified via 0G infrastructure) computes a real-time threat score using a robust multi-signal aggregation. Once the critical threshold is crossed, the system autonomously executes an emergency evacuation, leveraging 1inch to instantly route and convert volatile assets into a secure stablecoin. Designed with a transparent and aligned FinTech business model, Umbra is free to set up and charges a 0.5% success fee exclusively on the funds actually rescued from the crash, ensuring our incentives are 100% aligned with the user.
Umbra was built using a hybrid architecture combining a Python quantitative engine with a TypeScript execution layer. For data ingestion, we queried Uniswap v3 subgraphs via The Graph to stream real-time pool metrics and extract historical tick data.
We utilized 0G for decentralized data storage and to validate our custom quantitative risk model, the CBRI (CircuitBreaker Risk Index), ensuring our threat scoring mechanism remains immutable and transparent. For execution, we integrated 1inch as our production routing layer to seamlessly convert volatile assets into secure stablecoins at optimal rates during a market emergency.
To solve the limitation of historical DEX API quotes, our hackiest and most notable engineering feat was building an event-driven Python backtester from scratch during the hackathon. By replaying historical block data from the March 2023 USDC depeg and dynamically reconstructing liquidity depth from The Graph, we mathematically optimized our trigger thresholds using a Noisy-OR aggregation of liquidity drain velocity, pool imbalance, and price divergence, delivering a production-ready system.

