CCAR-F Practice Test 15 — 15 Questions

Question 1

A team is exposing an existing internal REST API to Claude through an MCP server. The API requires an OAuth 2.0 bearer token per user, and the team wants each user's Claude session to act with that user's own permissions. Where should the per-user OAuth token be handled in the MCP architecture?

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

Accepted Answer

Have the MCP server obtain and hold the user's OAuth token, attaching it to outbound API calls on the user's behalf.. Per-user authorization requires the MCP server to act as the trust boundary: it acquires each user's OAuth token through the client's authorization flow and attaches it to outbound API calls. This preserves least privilege and isolation while keeping credentials out of the model's context and the tool schema. Shared service accounts, schema-embedded tokens, and tokens returned in results all break the isolation or leak secrets.

Answer

Return the token to the model in the tool result and instruct the model to include it in subsequent calls.

Answer

Store a single shared service account token in the server's environment variables and use it for all users.

Answer

Embed the token in the tool's `inputSchema` so the model passes it as an argument on every call.

Question 2

A team is using the Anthropic Messages API to extract line items from vendor invoices. They define a tool named `record_line_items` with an input_schema that requires `description`, `quantity`, and `unit_price`. In production, Claude sometimes returns a text message describing the items instead of calling the tool, especially on invoices with unusual layouts. They want to guarantee a tool call every time. Which change is most effective?

Accepted Answer

Set `tool_choice` to `{"type": "tool", "name": "record_line_items"}` so the model must call that specific tool.. Anthropic's Messages API supports `tool_choice` with a named tool, which forces Claude to call that tool on every turn. This is the only mechanism that structurally guarantees the desired tool call rather than merely biasing the model toward it. Instructions, examples, and `tool_choice: any` all leave room for the model to respond with text or a different tool, so they cannot satisfy the 'every time' requirement.

Answer

Add few-shot examples showing Claude calling `record_line_items` for several invoice layouts.

Answer

Add a system prompt line instructing Claude to always call `record_line_items` and never respond with text.

Answer

Set `tool_choice` to `{"type": "any"}` so Claude can pick whichever tool it prefers.

Question 3

You are designing a prompt that asks Claude to review a pull request and return findings in a strict XML structure with a severity attribute for each finding. Your parser depends on the structure being consistent. Which TWO techniques most directly improve the consistency of the XML output? (Choose two.)

Accepted Answer

Describe the required tags and attributes precisely in the instructions, specifying allowed severity values.. Consistency in structured output comes from removing ambiguity. A concrete example shows the exact shape, and precise instructions define tags and allowed attribute values. Together they constrain the model to a stable format. Step-by-step reasoning, high temperature, and Markdown lists either do not enforce structure or actively introduce variation, so they do not meet the parser's needs in this scenario.

Answer

Ask the model to return the findings as a Markdown bulleted list for readability.

Answer

Set the temperature to a high value so the model explores different XML layouts.

Answer

Instruct the model to think step by step before producing the XML.

Question 4

An architect is reviewing an MCP server where one tool returns a 2 MB JSON payload of raw database rows. The model's responses have become slow and occasionally drop details from the middle of the payload. Which change best addresses the observed behavior while preserving the tool's usefulness?

Accepted Answer

Return a summarized or paginated subset of the rows, with the tool schema exposing parameters for filtering and page size.. Oversized tool results degrade latency and cause the model to lose details buried in low-signal data. Returning a summarized or paginated subset, with schema parameters that let the model filter and page, keeps each result within a size the model can attend to while preserving access to the full dataset on demand. Expanding context, compressing bytes, or deferring to a resource does not reduce the semantic volume the model must process.

Answer

Increase the model's context window and continue returning the full 2 MB payload on every call.

Answer

Compress the 2 MB payload with gzip and return the binary blob in the tool result.

Answer

Move the payload into an MCP resource and have the tool return only a resource URI for the model to fetch.

Question 5

An orchestrator is designed to handle 'Parallel Tool Use' for a travel booking agent. Which TWO conditions must be met for the architect to successfully implement this feature?

Accepted Answer

The tools being called must not have sequential dependencies on each other's outputs.. Parallel tool use allows Claude to suggest multiple actions in a single response, which can significantly reduce latency in complex workflows. To use this effectively, the architect must ensure the tools are independent of one another and that the application can handle and respond to multiple 'tool_use' IDs simultaneously.

Answer

The model must be instructed to use a single 'tool_result' block for all outputs.

Answer

The 'temperature' must be set to 0 to ensure the tools are called in alphabetical order.

Answer

A human must manually approve the parallel execution via a dedicated CLI tool.

Question 6

Refer to the exhibit. An architect is configuring an MCP server to provide read-only access to a SQLite database. What is the primary reason for using the 'env' field in this configuration?

Accepted Answer

To provide the necessary database credentials and enforce runtime constraints like read-only access.. The configuration of MCP servers involves defining how the host process should launch and communicate with the server. Using environment variables is a standard method for passing configuration and security constraints to the server process. This ensures that the server operates within the intended security boundaries, such as read-only mode, regardless of the model's requests.

Answer

To specify the transport protocol that the MCP host should use to communicate with the SQLite server.

Answer

To define the JSON schema for the tools that the SQLite server will expose to the model.

Answer

To increase the memory allocation for the 'uvx' command to handle larger database queries.

Question 7

When designing a multi-agent system, what is the primary benefit of the 'Orchestrator-Worker' pattern compared to a single-agent architecture with 20 tools?

Accepted Answer

It reduces the complexity of the orchestrator's context window by offloading details to workers.. As the number of tools increases, the complexity of the prompt and the likelihood of tool selection errors grow. The Orchestrator-Worker pattern simplifies the individual prompts by delegating specific tasks to specialized agents, which improves accuracy and makes the overall system easier to debug and maintain.

Answer

It eliminates the need for a system prompt for the worker agents.

Answer

It allows the model to bypass the per-token pricing of the Anthropic API.

Answer

It guarantees that the model will never produce a hallucinated tool name.

Question 8

A company is deploying an agent that interacts with a sensitive financial database. To maintain security and accuracy, the architect wants to ensure that any 'delete' or 'transfer' actions are reviewed by a human before execution. Which agentic pattern should be used?

Accepted Answer

An 'Interrupt-and-Approve' orchestration loop for specific tool definitions.. Human-in-the-loop (HITL) is a critical design pattern for agentic systems operating in high-stakes environments. It allows the model to propose actions while reserving the actual execution for a human-verified trigger. This mitigates the risk of autonomous errors and ensures that the agent operates within the safety guardrails of the organization.

Answer

Parallel Tool Execution to allow the model to run multiple checks simultaneously.

Answer

The ReAct pattern to allow the model to think before it acts on the database.

Answer

Increasing the frequency of system prompt injections to remind the model of the rules.

Question 9

When implementing the 'Evaluator-Optimizer' pattern for code generation with Claude 3.5 Sonnet, which TWO components are essential for the feedback loop to be effective?

Accepted Answer

A distinct set of evaluation criteria or unit tests for the code. The Evaluator-Optimizer pattern relies on a clear separation of generation and critique. To be effective, the evaluator must have a set of objective criteria or tests to check the generator's work, and the generator must receive the specific results of that evaluation to make targeted improvements in the next iteration of the loop.

Answer

The use of two different model families to ensure diversity of thought

Answer

Setting the temperature to 0 for both the generator and the evaluator

Answer

A mandatory human approval step after every single model iteration

Question 10

Refer to the exhibit showing a configuration for Claude 3.5 Sonnet. Which architectural requirement is strictly necessary when implementing this 'computer_20241022' tool in a production environment?

Accepted Answer

The tool must be executed within a dedicated, isolated sandbox or virtual machine. The 'computer_20241022' tool enables Claude to interact with a graphical user interface by taking screenshots and simulating inputs. Because the model essentially 'sees' the screen, the architect must provide a secure, isolated sandbox environment where the agent can execute commands safely without risking the host system or accessing sensitive data outside the intended scope.

Answer

The agent must have direct root access to the host operating system

Answer

The orchestration layer must include a visual feedback loop for human monitoring

Answer

The 'display_width_px' must always match the resolution of the developer's monitor

Question 11

Which TWO practices are recommended when defining tool schemas (JSON descriptions) for an agentic system to improve reliability and reduce errors?

Accepted Answer

Providing detailed descriptions for every parameter in the schema. Well-defined tool schemas are the foundation of effective agentic behavior. Providing clear, descriptive names and detailed property descriptions helps the model understand exactly when and how to use a tool, reducing the likelihood of incorrect parameter values or inappropriate tool selection during complex tasks.

Answer

Using short, cryptic tool names to save input tokens

Answer

Always marking all parameters as 'required' to ensure data integrity

Answer

Omitting the 'description' field for the tool itself to allow for flexibility

Question 12

An architect is designing an agentic system to handle complex software engineering tasks. The system must break down a high-level requirement into specific coding milestones, assign those milestones to specialized sub-agents, and synthesize the results. Which orchestration strategy is most appropriate?

Accepted Answer

The Orchestrator-Worker pattern. The Orchestrator-Worker pattern is specifically designed for complex, non-deterministic tasks where the central model must dynamically plan and delegate work. Unlike a simple router, the orchestrator monitors the progress of sub-tasks and synthesizes the outputs into a coherent final result, making it the superior choice for multi-step engineering projects.

Answer

The Chain-of-Thought Sequential pattern

Answer

The Router pattern

Answer

The Parallel Voting pattern

Question 13

When designing an MCP server, what is the best practice for handling dependencies between multiple tools?

Accepted Answer

Design each tool to be atomic and rely on the model to orchestrate the sequential execution.. The best practice is to design tools to be stateless and independent, requiring the orchestration logic to reside within the LLM agent or client. By ensuring that each tool performs a single, atomic action, you minimize coupling. This approach makes the tools easier to test, reuse, and maintain, and prevents complex state-dependency issues that can occur when one tool relies on the side effects of another in the same session.

Answer

Chain tool calls by storing state in a global variable within the MCP server process.

Answer

Combine dependent tools into a single, monolithic tool to guarantee the order of execution.

Answer

Force the model to provide a 'chain_id' parameter to track the context across multiple tool calls.

Question 14

What is the primary advantage of using a dedicated MCP server over embedding tool logic directly into the application code?

Accepted Answer

It enables independent scaling and easier maintenance of the tool logic.. Separating the MCP server allows for independent scaling, improved security through isolated execution environments, and modularity. By moving tools into their own processes, the architect can maintain, update, and secure these tools without redeploying the entire application. This separation of concerns is fundamental for building large-scale agentic systems where tool sets evolve at different speeds than the main application, ensuring stability and ease of integration.

Answer

It eliminates the need for any tool documentation or description.

Answer

It forces the model to use the tools more frequently.

Answer

It removes the need to validate inputs as the server is 'trusted'.

Question 15

To reduce the risk of Claude hallucinating information when it doesn't know the answer, what instruction should be added to the prompt?

Accepted Answer

Tell the model to say 'I don't know' if unsure.. Explicitly giving the model an 'out' is a highly effective way to prevent hallucinations. By instructing Claude to say 'I don't know' or 'Information not found' when it is uncertain, you steer the model away from its natural tendency to be helpful by generating plausible but incorrect answers.

Answer

Tell the model to 'Be as creative as possible'.

Answer

Instruct the model to 'Always provide an answer'.

Answer

Set the max_tokens to a very high number.