Build a Meshery MCP Tool - Exam

Question 1

Why build the first MCP tool as read-only?

Select one answer

Read-only tools cannot be logged

It is the safe foundation - observe before you allow changes

MCP does not support writes

It avoids needing a schema

Question 2

What makes a good MCP tool definition? (Select all that apply.)

Select all that apply

A clear, single-responsibility name

Typed inputs and a shaped output

Structured errors instead of crashing

One tool that does everything with write access

Question 3

What typically backs a Meshery MCP tool?

Select one answer

Meshery's REST/GraphQL API (or kubectl)

A spreadsheet

The model's training data

A container registry

Question 4

In MCP, what is a 'resource' (as opposed to a tool)?

Select one answer

Read-only data the agent can fetch on demand

A mutating action

A Kubernetes CRD

A load generator

Question 5

What should you do before connecting the MCP server to an agent?

Select one answer

Test it in isolation and check its logs

Give it write access to production

Remove its error handling

Skip testing to save time

Question 6

Which Meshery component would a resource exposing live cluster state read from, and why note its lag?

Type your answer below