Introducing MCP Safety Warden: a proxy for vetting MCP servers and enabling safer tool execution

Hey everyone I’ve been experimenting with MCP security and built an early-stage framework/server called MCP Safety Warden. The idea is to place a proxy layer between agents and MCP servers so tools are not blindly trusted before execution. It focuses on vetting MCP servers, profiling tool behavior,