"Vocabulary: Agents, Tools, and MCP"

1. What is the primary function of a tool schema in AI models?

To request actions from the runtime To declare the model’s capabilities To define machine-parseable output To validate requests before execution

2. What is meant by 'least-privilege tools' in the context of AI security?

Tools that operate with maximum capabilities Tools that require no human confirmation Tools with limited permissions necessary for their tasks Tools that allow full database access without restrictions

3. In what manner does the Model Context Protocol (MCP) benefit AI integration?

By limiting the capabilities of AI clients By providing a standard way for clients to discover tools By enforcing strict permissions on data access By controlling data input from users

4. What role does 'human confirmation gate' play in AI tool execution?

It allows the model to bypass runtime security It requires human approval for consequential actions It ensures tools operate independently of any oversight It records transaction logs for all actions taken

Continue: "Vocabulary: MLOps and Evaluation" → Review this lesson Retake quiz