Open-source enforcement at runtime that masks sensitive data and blocks destructive commands from AI agents. You change one thing: your agent's port.
Open on GitHub