CCARF Domain D2
Tool Design & MCP Integration Practice Questions
About 18% of the exam. MCP tool/resource design, clear tool descriptions, structured errors (isError), .mcp.json scoping, and choosing built-in tools (Read, Write, Edit, Bash, Grep, Glob).
36 practice questions · Official blueprint weight 18% · Claude Certified Architect Foundations (CCAR-F)
Paper 1 · Q2A developer productivity platform has a tool called analyze_content with the description 'Analyses content from various sources.' Logs show the agent uses it fo…
Paper 1 · Q5An agent has two tools: `analyze_content` ('Analyses content') and `extract_web_results` ('Extracts data from web pages'). After a system prompt update added 'A…
Paper 1 · Q20A support agent has `process_refund` and `adjust_billing` tools with clearly differentiated descriptions, yet still occasionally calls `adjust_billing` for refu…
Paper 1 · Q21A developer needs to rename a variable from 'userData' to 'customerData' in a specific file. The variable appears 12 times throughout the file. What is the most…
Paper 1 · Q23A customer support team configures an MCP server for their internal CRM system. The agent frequently ignores the CRM MCP tool and instead uses the built-in Grep…
Paper 1 · Q27A developer productivity platform has a code review agent equipped with 18 tools covering code analysis, documentation lookup, test generation, dependency check…
Paper 1 · Q33A platform has `search_knowledge_base` ('Searches help articles') and `process_action` ('Handles customer actions like refunds and plan changes'). The agent pic…
Paper 1 · Q36A team needs to integrate with Jira for issue tracking in their Claude Code workflow. A developer proposes building a custom MCP server. What is the correct fir…
Paper 1 · Q41A team's agent must generate a weekly report on the first turn of each conversation by calling `generate_report`. With `tool_choice: 'auto'`, the agent sometime…
Paper 1 · Q50A customer support agent has two tools: get_customer_profile (returns name, email, preferences) and get_customer_orders (returns order history with statuses). U…
Paper 1 · Q59An engineering team needs to update all references to a renamed API endpoint across their 500,000-line codebase. They know the old endpoint name (POST /api/v1/u…
Paper 2 · Q2The data platform team has built an MCP server with 22 tools: query_snowflake, query_postgres, query_api, plus 19 specialised tools for individual data transfor…
Paper 2 · Q10A `query_postgres` MCP tool returns 500 on transient server overload and 404 on missing tables. The agent retries both identically, exhausting retries on missin…
Paper 2 · Q15An enterprise data platform agent has 22 tools available, including query tools for Snowflake, PostgreSQL, and third-party APIs, plus data transformation, visua…
Paper 3 · Q6[Scenario 1: Customer Support Resolution Agent] You are building a customer support resolution agent with the Claude Agent SDK. It handles high-ambiguity reques…
Paper 3 · Q7[Scenario 1: Customer Support Resolution Agent] You are building a customer support resolution agent with the Claude Agent SDK. It handles high-ambiguity reques…
Paper 3 · Q8[Scenario 1: Customer Support Resolution Agent] You are building a customer support resolution agent with the Claude Agent SDK. It handles high-ambiguity reques…
Paper 3 · Q26[Scenario 3: Multi-Agent Research System] A coordinator agent built on the Claude Agent SDK delegates to specialized subagents — web search, document analysis, …
Paper 3 · Q27[Scenario 3: Multi-Agent Research System] A coordinator agent built on the Claude Agent SDK delegates to specialized subagents — web search, document analysis, …
Paper 3 · Q31[Scenario 4: Developer Productivity with Claude] You are building developer productivity tooling on the Claude Agent SDK: helping engineers explore unfamiliar c…
Paper 3 · Q32[Scenario 4: Developer Productivity with Claude] You are building developer productivity tooling on the Claude Agent SDK: helping engineers explore unfamiliar c…
Paper 3 · Q33[Scenario 4: Developer Productivity with Claude] You are building developer productivity tooling on the Claude Agent SDK: helping engineers explore unfamiliar c…
Paper 3 · Q34[Scenario 4: Developer Productivity with Claude] You are building developer productivity tooling on the Claude Agent SDK: helping engineers explore unfamiliar c…
Paper 3 · Q57[Scenario 6: Structured Data Extraction] You are building a structured data extraction system: pulling information from unstructured documents, validating outpu…
Paper 3 · Q58[Scenario 6: Structured Data Extraction] You are building a structured data extraction system: pulling information from unstructured documents, validating outpu…
Paper 4 · Q2A synthesis agent frequently returns control to the coordinator for simple fact verification, adding 2–3 round trips per task and 40% latency. Analysis shows 85…
Paper 4 · Q9A developer is investigating a production bug. They know the error message 'InvalidStateTransition' is logged somewhere in the codebase, but they do not know wh…
Paper 4 · Q14Production logs show an agent frequently calls get_customer when users ask about orders (e.g. 'check my order #12345'), instead of calling lookup_order. Both to…
Paper 4 · Q20A research subagent's database query fails partway through an investigation. Which pieces of information belong in the structured error context it returns to th…
Paper 4 · Q21A tool returns an empty array after a customer lookup. The agent retries 3 times, then escalates to a human agent. Analysis shows the customer's account simply …
Paper 4 · Q25A team's shared `.mcp.json` configures a PostgreSQL MCP server. A developer adds a personal staging database MCP server to their `~/.claude.json`, so both datab…
Paper 5 · Q7[Scenario: Multi-agent Research System] Production logs show a persistent pattern: requests like “analyze the uploaded quarterly report” are routed to the web-s…
Paper 5 · Q9[Scenario: Multi-agent Research System] In your system design, you gave the document analysis agent access to a general-purpose tool `fetch_url` so it could dow…
Paper 5 · Q38[Scenario: Customer Support Agent] Production logs show a consistent pattern: when customers include the word “account” in their message (e.g., “I want to check…
Paper 5 · Q40[Scenario: Conversational AI Architecture Patterns] Your `remove_team_member` tool uses a `dry_run: boolean` parameter for previewing impacts before execution. …
Paper 5 · Q41[Scenario: Conversational AI Architecture Patterns] Production monitoring shows your `search_catalog` tool fails 12% of the time: 8% are network timeouts that s…
Other domains
D1 Agentic Architecture & Orchestration · D3 Claude Code Configuration & Workflows · D4 Prompt Engineering & Structured Output · D5 Context Management & Reliability