AnthropicTool Design and MCP Integration

Mcp & Trust Boundary — CCAR-F Practice Question

A representative Claude Certified Architect - Foundations (CCAR-F) exam question on Mcp & Trust Boundary. Work through it below, then read why each option is right or wrong.

Short answer

The correct answer is B. Server output is untrusted; scope credentials, validate responses, delimit content, and restrict exposed capabilities.

MCP standardizes connectivity, not trust. Third-party output can be malicious or malformed, so normal authentication, validation, and prompt-injection defenses still apply.

The Question

A remote MCP server is operated by a third party and returns content that enters the model context. What assumption should the architecture make?

AAll server content is trusted because it uses MCP
BServer output is untrusted; scope credentials, validate responses, delimit content, and restrict exposed capabilitiesCorrect
CMCP removes the need for authentication
DThe model can identify every malicious instruction by itself

Why B is correct

MCP standardizes connectivity, not trust. Third-party output can be malicious or malformed, so normal authentication, validation, and prompt-injection defenses still apply.

Why the other options are wrong

Option A: All server content is trusted because it uses MCP

Option A does not satisfy the requirement in the scenario. Review the explanation above: the correct choice (B) is the only one that fully meets every constraint stated in the question.

Option C: MCP removes the need for authentication

Option C does not satisfy the requirement in the scenario. Review the explanation above: the correct choice (B) is the only one that fully meets every constraint stated in the question.

Option D: The model can identify every malicious instruction by itself

Option D does not satisfy the requirement in the scenario. Review the explanation above: the correct choice (B) is the only one that fully meets every constraint stated in the question.

Key idea: Mcp & Trust Boundary

MCP standardizes connectivity, not trust. Third-party output can be malicious or malformed, so normal authentication, validation, and prompt-injection defenses still apply. On the CCAR-F exam, questions in the "Tool Design and MCP Integration" domain test whether you can map a scenario's constraints to the right choice. Read the requirement carefully, eliminate options that violate any single constraint, and pick the one that satisfies all of them with the least operational overhead.

Ready to see how you'd score?

Take the free practice quiz and find out which Claude Certified Architect - Foundations domains you need to focus on. No signup required.

Practice 5 similar questions

Same cert, same or adjacent domain. Use these after reviewing the explanation.

Related Claude Certified Architect - Foundations Practice Questions