YieldSentinel

Problem Statement

DeFi agents lack privacy-preserving infrastructure and safe yield management tooling. Current agents expose user strategies through centralized inference providers, and there's no standard tooling for AI agents to safely interact with Lido � staking, monitoring positions, or spending yield without risking principal. We're building an MCP server for Lido with full stETH/wstETH integration, vault monitoring, and yield-only spending controls, all routed through Venice for private inference.