Emerging protocols like Anthropic’s MCP and Google Cloud’s A2A introduce a new dimension of machine-to-machine interaction between agentic workflows. These workflows rely on Non-Human Identities (NHIs) like API keys, OAuth apps, and service accounts to function—identities that are often highly privileged and under-monitored. As AI-to-AI communication scales, so does the risk of ungoverned identity sprawl and unauthorized access paths within critical systems. Understanding how these protocols work—and the identity implications they introduce—is essential for building secure, future-ready architectures.