CCDV-F Tools and MCP Integration — 25 Questions

Question 1

When designing an MCP Server to help Claude interact with a local database, a developer needs to decide between using MCP Resources or MCP Tools. Which TWO statements correctly describe the differences between these primitives?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Resources are intended for read-only data that Claude can use as context.. Choosing the right MCP primitive is essential for optimal model performance. Resources are passive data sources that Claude can read for context, similar to documentation or files. Tools are active components that allow the model to perform actions or execute code. Distinguishing between data retrieval and action execution ensures that the model uses the most efficient path for information.

Answer

Resources require a JSON schema for input arguments just like tools do.

Answer

Tools are automatically called by the MCP Host without model intervention.

Answer

Resources are always updated in real-time via a WebSocket push mechanism.

Question 2

A developer wants to ensure that Claude uses a specific 'search_documentation' tool every single time a user asks a question about their proprietary software. Which configuration of the 'tool_choice' parameter should be used?

Accepted Answer

Set tool_choice to {'type': 'tool', 'name': 'search_documentation'}.. The 'tool_choice' parameter allows developers to control Claude's tool-calling behavior. By default, 'auto' lets Claude decide if a tool is needed. Setting 'tool_choice' to 'tool' with a specific name forces Claude to use that tool for its next response, regardless of whether it thinks it needs it, which is ideal for rigid workflows.

Answer

Set tool_choice to 'auto' and include the instruction in the system prompt.

Answer

Set tool_choice to 'any' to allow Claude to pick any available tool.

Answer

Omit the tool_choice parameter entirely as it defaults to 'required'.

Question 3

A developer is building a complex MCP-based system. They want to utilize 'MCP Prompts' to standardize how Claude interacts with their server. Which THREE statements accurately describe the capabilities and behavior of MCP Prompts?

Accepted Answer

They allow servers to expose reusable templates with variable arguments.. MCP Prompts are a powerful feature that allows servers to provide predefined templates for the model. They help standardize interactions and can be used to guide the model through complex tasks. Unlike tools or resources, prompts are specifically designed to influence the model's instructions and conversation structure directly from the server side.

Answer

They are automatically injected into the system prompt of every request.

Answer

Prompts are executed as Python code on the server before being sent.

Question 4

Which TWO of the following are benefits of using the Model Context Protocol (MCP) over traditional, custom-built tool-calling implementations?

Accepted Answer

It allows a single server to be used across different host applications like Claude Desktop and IDEs.. MCP provides a standardized interface that decouples the model from the data sources. This standardization reduces the effort needed to integrate new tools and allows for a more modular ecosystem. By using a common protocol, developers can swap out different models or hosts without rewriting their entire integration layer for every tool.

Answer

It automatically encrypts all local data before sending it to the Anthropic API.

Answer

It reduces the need for the model to reason about which tool to use for a task.

Answer

It eliminates the latency associated with network-based API calls.

Question 5

When providing a system prompt alongside a set of tools, what is the best practice for instructing Claude on how to use those tools?

Accepted Answer

Focus the system prompt on the model's persona and high-level goals.. While Claude is trained to use tools based on their JSON definitions, the system prompt can provide critical context on *when* and *why* to use them. Effective system prompts guide the model's strategy, helping it understand complex business logic or multi-step processes that are not apparent from the individual tool descriptions alone.

Answer

Repeat the entire JSON schema of every tool inside the system prompt.

Answer

Tell the model to ignore the tool descriptions and follow the prompt instead.

Answer

Use the system prompt to provide the API keys needed to run the tools.

Question 6

A developer wants to ensure that Claude always uses a specific tool named 'search_documentation' to answer any user query, regardless of the user's intent. Which TWO configurations or strategies are necessary to achieve this? (Select TWO)

Accepted Answer

Set the tool_choice type to 'tool' and provide the name 'search_documentation'.. Forcing tool use is a common pattern for specialized agents. This is achieved by setting the 'tool_choice' parameter to 'tool' and specifying the exact name of the tool. This configuration overrides Claude's default behavior of deciding whether a tool is necessary, ensuring that the defined tool is invoked in every turn of the conversation.

Answer

Set the tool_choice type to 'auto' and include only one tool in the list.

Answer

Set the tool_choice type to 'any' to force the selection of at least one tool.

Answer

Omit the 'description' field in the tool definition to prevent model confusion.

Question 7

The Model Context Protocol (MCP) introduces a standardized way to connect models to data and tools. Which TWO statements accurately describe the roles of MCP Clients and Servers? (Select TWO)

Accepted Answer

An MCP Server can provide tools, resources, and prompt templates to a client.. MCP clarifies the relationship between the model and external resources. Servers are the providers of capabilities, while clients are the consumers that manage the model's interaction with those capabilities. This standardized interface allows a single MCP server to be used across multiple different AI applications without rewriting the tool integration logic.

Answer

The MCP Client is responsible for executing the Python code defined in the server.

Answer

MCP Servers must be hosted on public cloud infrastructure to be accessible.

Answer

The MCP Client defines the JSON schema for all tools provided by the server.

Question 8

A developer is building a Claude-powered support assistant that must query an internal ticketing system through a custom MCP server named 'tickets-mcp'. The server exposes a tool called 'get_ticket' and is launched as a local stdio process. During integration, Claude never emits a 'tool_use' block for 'get_ticket', even when the user explicitly asks about a ticket number. The developer has verified that the tool schema is valid JSON and the server starts without errors. Which configuration issue is the MOST likely cause?

Accepted Answer

The developer forgot to include the MCP server's tool definitions in the 'tools' parameter of the Messages API request, so Claude has no knowledge of 'get_ticket'.. The strongest explanation is that the MCP server's tools were never forwarded into the Messages API 'tools' parameter. MCP standardizes discovery and invocation, but the model only sees tools that the client includes in the request. A healthy server with a valid schema can still produce zero tool_use blocks if its definitions are not passed through, which precisely matches the silent failure in this scenario.

Answer

The stdio server must be started with the '--expose-tools' flag before Claude can invoke any tool it defines.

Answer

The 'tickets-mcp' server was configured with the 'resource' capability instead of the 'tools' capability, so 'get_ticket' is not advertised to Claude.

Answer

The 'get_ticket' tool description exceeds 1024 characters, causing Claude to silently ignore the tool during planning.

Question 9

A developer is implementing a weather-reporting agent using Claude 3.5 Sonnet. After Claude generates a tool_use block for the 'get_weather' tool, the developer fetches the data. What is the correct next step to ensure Claude can finalize its response to the user?

Accepted Answer

Send a new user message containing a 'tool_result' block with the 'tool_use_id'.. The tool use lifecycle requires a multi-turn interaction where the model first requests a tool execution. Once the developer executes the tool locally, they must return the result to Claude in a new 'user' message with a 'tool_result' block. This allows Claude to incorporate the external data into its final synthesis and provide an accurate answer to the original query.

Answer

Append the tool results directly to the end of the previous assistant message.

Answer

Update the system prompt with the new weather data and restart the session.

Answer

Issue a second assistant message that includes the weather data as a JSON string.

Question 10

In the Model Context Protocol (MCP) architecture, which component is responsible for providing specific resources, tools, and prompts to the rest of the ecosystem?

Accepted Answer

The MCP Server. MCP is designed as a client-server architecture to standardize how AI models access external data and functions. The MCP Server is the provider in this relationship, exposing specific capabilities such as local files, database schemas, or API integrations. Understanding this role is fundamental for developers building custom integrations that extend Claude's capabilities beyond its training data.

Answer

The MCP Host

Answer

The MCP Client

Answer

The MCP Gateway

Question 11

Refer to the exhibit. A developer provides this tool definition to Claude. If a user asks 'What is the tax on $100?', how will Claude likely behave based on the provided JSON schema?

Accepted Answer

Claude will likely ask the user which state they are in before calling the tool.. The JSON schema defines the contract between the model and the tool. In this exhibit, 'amount' is required, but 'state_code' is not. However, the description implies 'state_code' is necessary for an accurate calculation. Claude will attempt to follow the schema first, but if it identifies that a critical piece of information for the logic is missing, it may ask for clarification.

Answer

Claude will trigger an API error because the state_code is missing from the required list.

Answer

Claude will call the tool with a default state_code of 'CA'.

Answer

Claude will automatically extract the state from the user's IP address.

Question 12

An MCP Server is connected to a Host using the Stdio transport. If the server process crashes, how does the MCP architecture generally handle the reconnection logic?

Accepted Answer

The Host application detects the closed stream and must manage the restart.. The Stdio transport relies on standard input and output streams for communication between the host and the server. Because this is a process-level connection, a crash usually results in the termination of the stream. The responsibility for detecting this failure and initiating a recovery or restart lies with the MCP Host application.

Answer

The MCP Client sends a 'heartbeat' signal to automatically restart the server.

Answer

The MCP Server uses a sidecar process to monitor its own health and restart.

Answer

The Model will identify the crash and suggest a fix to the developer.

Question 13

Claude requests a tool call to 'query_database', but the database is currently down. What is the best practice for handling this error so Claude can inform the user effectively?

Accepted Answer

Send a 'tool_result' with is_error: true and a descriptive error message.. Graceful error handling is vital for a good user experience. Instead of failing the entire API request, the developer should return the error message within the 'tool_result' block. By setting 'is_error' to true, the developer signals to Claude that the tool failed, allowing the model to explain the situation to the user.

Answer

Return a 500 Internal Server Error to the Anthropic API request.

Answer

Omit the 'tool_result' and send a new user message saying 'The tool failed'.

Answer

Modify the assistant's previous tool_use block to remove the request.

Question 14

Refer to the exhibit. A developer is debugging an MCP integration and sees these logs in their console. What is the most likely cause of this error?

Accepted Answer

The 'get_files' tool is not correctly defined in the server's list_tools handler.. The error code -32601 is a standard JSON-RPC error indicating that the requested method does not exist on the server. In the context of MCP, this means the Host is trying to call a tool named 'get_files', but the Server has not registered or implemented a tool with that exact name in its tool list.

Answer

The MCP Server is using an outdated version of the protocol.

Answer

The path argument '/' is restricted by the server's security policy.

Answer

The model (Claude) generated a malformed JSON object for the arguments.

Question 15

When using Claude's tool use features in a streaming response, how should the developer handle the 'tool_use' blocks in the stream?

Accepted Answer

Accumulate tool_use delta events until a 'content_block_stop' event is received.. Streaming with tool use requires careful handling of partial blocks. As Claude streams its response, it may start a 'tool_use' block. The developer must accumulate these fragments until the block is complete before attempting to execute the tool, as the arguments and even the tool name may be delivered across multiple chunks.

Answer

Execute the tool as soon as the 'name' property is received in the stream.

Answer

Wait for the 'message_stop' event before parsing any tool_use blocks.

Answer

Streaming is not supported when using tools; use the synchronous API instead.

Question 16

Claude generates a response containing three separate 'tool_use' blocks in a single turn. How should the developer process these tool calls to maintain optimal performance and follow Anthropic's best practices?

Accepted Answer

Execute all tools in parallel and return all results in one user message.. Claude is capable of requesting multiple tool calls simultaneously if the user's request requires it. To minimize latency, developers should execute these calls in parallel whenever possible. Once all results are gathered, they must be returned to Claude in the same sequence and within a single user message to continue the conversation.

Answer

Process each tool call sequentially to avoid race conditions in the model.

Answer

Only execute the first tool call and ignore the others to prevent token bloat.

Answer

Ask the user to prioritize which of the three tools should be run first.

Question 17

A security auditor is reviewing an application that uses the Model Context Protocol (MCP). Which TWO security considerations are most important for the developer to address when implementing the MCP Server?

Accepted Answer

The server should strictly validate all tool arguments against its internal logic.. Security in MCP is a shared responsibility. While the protocol facilitates communication, the server developer must ensure that the tools and resources exposed do not create vulnerabilities. This includes limiting access to the file system and ensuring that tool inputs are properly validated before being used in potentially dangerous operations like shell commands.

Answer

The server must use a unique API key for every tool call it receives.

Answer

The server must rotate its connection ID every sixty seconds for privacy.

Answer

All MCP tool results must be manually approved by a human operator.

Question 18

A company wants Claude to answer questions about their internal HR policies. They are deciding between fine-tuning a model and using a Tool-calling (RAG) approach. Why is Tool-calling usually preferred for this use case?

Accepted Answer

Tool-calling allows the model to access the most up-to-date policies instantly.. Tool-calling, often used in Retrieval-Augmented Generation (RAG), is generally superior for internal knowledge bases because it allows for easy updates and provides verifiable citations. Fine-tuning is expensive, time-consuming, and results in a 'frozen' snapshot of data that becomes obsolete as soon as policies change, making it unsuitable for dynamic corporate environments.

Answer

Fine-tuning cannot be used to teach a model new factual information.

Answer

Tool-calling is the only way to ensure Claude uses a professional tone.

Answer

Fine-tuned models are restricted from using tools for security reasons.

Question 19

A developer is architecting a system where Claude must interact with a proprietary internal database. According to the Model Context Protocol (MCP) architecture, which component is responsible for surfacing the database schemas and handling the actual execution of the SQL queries?

Accepted Answer

MCP Server. In the Model Context Protocol architecture, the MCP Server is the specific component designed to host tools and resources. It acts as the interface to local or remote data sources, providing the necessary context and execution environment. Understanding this separation of concerns is vital for developers to ensure that security boundaries and data access are managed correctly within the MCP ecosystem.

Answer

MCP Client

Answer

MCP Transport

Answer

MCP Host

Question 20

Refer to the exhibit. A developer provides this tool definition to Claude. If the user asks for the weather in 'London' without specifying a unit, how will Claude's tool_use block be structured based on this JSON schema?

Accepted Answer

It will include only the 'location' parameter in the input object.. Claude follows the provided JSON schema strictly when generating tool calls. Since the 'unit' property is not included in the 'required' array, Claude is not obligated to provide it if the user does not specify it. This behavior demonstrates Claude's ability to handle optional parameters while adhering to the structure defined in the tool's input_schema field.

Answer

It will fail with a validation error because 'unit' is missing.

Answer

It will prompt the user to choose between 'celsius' and 'fahrenheit' first.

Answer

It will default the 'unit' to 'celsius' automatically.