Three Moons Lab is an independent AI infrastructure lab focused on building practical tools for agent reliability, release readiness, and governance
Three Moons Lab is an independent AI infrastructure lab focused on building practical tools for agent reliability, release readiness, and governance. Our current work explores how AI agents can safely use tools, interact with external systems, and ship with higher confidence. We build developer-native infrastructure for evaluating tool surfaces, identifying release risks, and helping teams reason about agent readiness before production deployment. Current project: agents-shipgate — an OpenAI Agents SDK-first release readiness tool focused on tool-use readiness for MCP and OpenAPI-based agent systems. https://github.com/ThreeMoonsLab/agents-shipgate