CCARF

CCARF Domain D5

Context Management & Reliability Practice Questions

About 15% of the exam. Context budgets, progressive summarization risks, lost-in-the-middle effects, escalation criteria, structured error propagation, and human-review calibration.

43 practice questions · Official blueprint weight 15% · Claude Certified Architect Foundations (CCAR-F)

Paper 1 · Q3Claude Code is auditing API documentation coverage across a large codebase. After exploring 15 modules, the agent produces vague references like 'the authentica… Paper 1 · Q6A data extraction pipeline extracts financial figures from annual reports. The pipeline cites sources using inline text references like 'According to page 12 of… Paper 1 · Q7A docs team processes a 500,000-line codebase by having Claude Code read files and generate docs, instructing it to 'summarise each module after documenting it,… Paper 1 · Q9A customer asks a support agent to match a lower price advertised by a competitor. The agent's policy documents cover price adjustments for the company's own si… Paper 1 · Q16A legal document extraction pipeline has been running for three months with calibrated confidence thresholds. Extractions above 90% confidence are auto-approved… Paper 1 · Q24A documentation team working across a 500,000-line codebase notices that Claude Code's responses become less specific after extensive file exploration. What is … Paper 1 · Q29A documentation generation pipeline uses Claude Code to produce docs from three sources: source code comments, existing wiki pages, and API schema files. During… Paper 1 · Q31A structured data extraction system processes invoices, purchase orders, and contracts. The monitoring dashboard shows 97% overall extraction accuracy and the t… Paper 1 · Q40A customer support agent achieves only 55% first-contact resolution, well below the 80% target. Logs show it escalates straightforward damage replacement cases … Paper 1 · Q49Claude Code is synthesising release notes from commit messages, PR descriptions, and changelog entries. During a 200+ commit session, summaries of early commits… Paper 1 · Q55Your extraction pipeline reports 97% aggregate accuracy, and you want to automate high- confidence extractions. Which safeguards should be in place before you r… Paper 2 · Q6A data engineer delegates a complex cross-database analysis to the platform agent: join Snowflake billing data with PostgreSQL usage data, apply currency normal… Paper 2 · Q11An agent receives 200 rows of Snowflake financial data and places it mid-prompt, between system instructions and the user's follow-up question. Asked to identif… Paper 3 · Q9[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 · Q10[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 · Q19[Scenario 2: Code Generation with Claude Code] Your team uses Claude Code for code generation, refactoring, debugging, and documentation, integrated into daily … Paper 3 · Q20[Scenario 2: Code Generation with Claude Code] Your team uses Claude Code for code generation, refactoring, debugging, and documentation, integrated into daily … Paper 3 · Q28[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 · Q29[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 · Q30[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 · Q59[Scenario 6: Structured Data Extraction] You are building a structured data extraction system: pulling information from unstructured documents, validating outpu… Paper 3 · Q60[Scenario 6: Structured Data Extraction] You are building a structured data extraction system: pulling information from unstructured documents, validating outpu… Paper 4 · Q1A multi-agent system has subagents for web search, academic database, and industry reports. The academic database subagent hits a 403 Forbidden on a specific jo… Paper 4 · Q4A customer support agent handles a multi-issue session. After several turns, the agent refers to 'your recent refund request' instead of the specific $247.83 re… Paper 4 · Q5A research subagent queries an academic journal database for articles on renewable energy policy. The database responds successfully but returns zero matching a… Paper 4 · Q8A customer says: 'This is ridiculous, I have been waiting 20 minutes. Just connect me to a real person.' The agent has access to the customer's account and can … Paper 4 · Q10A multi-agent research system has three subagents (financial filings, news, technical white papers) and a synthesis agent. Each subagent returns properly attrib… Paper 4 · Q19A customer support agent searches for a customer named 'Sarah Johnson' and the lookup tool returns three matching accounts with the same name but different addr… Paper 4 · Q26A multi-agent research system finds conflicting data on a pharmaceutical compound's efficacy. A 2022 peer-reviewed clinical trial reports 78% efficacy, while a … Paper 4 · Q33A customer support agent uses a token budget of 200k tokens. The system prompt consumes 8k tokens, conversation history takes 120k tokens, and the most recent t… Paper 4 · Q37A web search subagent in a multi-agent research system times out while researching a complex topic. You need to design how this failure information flows back t… Paper 5 · Q3[Scenario: Multi-agent Research System] A document analysis subagent frequently fails when processing PDF files: some have corrupted sections that trigger parsi… Paper 5 · Q5[Scenario: Multi-agent Research System] The web-search subagent returns results for only 3 of 5 requested source categories (competitor sites and industry repor… Paper 5 · Q6[Scenario: Multi-agent Research System] The document analysis subagent encounters a corrupted PDF file that it cannot parse. When designing the system’s error h… Paper 5 · Q11[Scenario: Multi-agent Research System] Production monitoring shows inconsistent synthesis quality. When aggregated results are ~75K tokens, the synthesis agent… Paper 5 · Q12[Scenario: Multi-agent Research System] In testing, the combined output of the web-search agent (85K tokens including page content) and the document analysis ag… Paper 5 · Q33[Scenario: Code Generation with Claude Code] You’re adding error-handling wrappers around external API calls across a 120-file codebase. The work has three phas… Paper 5 · Q43[Scenario: Conversational AI Architecture Patterns] Users refine playlist preferences over multiple conversation turns. Two messages after a user said "I love j… Paper 5 · Q44[Scenario: Conversational AI Architecture Patterns] After a 40-minute cooking session, the conversation reaches 78,000 tokens. History includes allergies, recip… Paper 5 · Q45[Scenario: Conversational AI Architecture Patterns] Users report that during extended conversations the assistant loses track of earlier topics and preferences.… Paper 5 · Q46[Scenario: Conversational AI Architecture Patterns] Users report that latency increases and costs rise when conversations exceed 50 turns. What is the primary c… Paper 5 · Q47[Scenario: Conversational AI Architecture Patterns] After three months of weekly sessions, conversation history grows to 85,000 tokens. When a user asks "What d… Paper 5 · Q52[Scenario: Conversational AI Architecture Patterns] A webhook notifies your system that a user's package has shipped while the user is actively chatting. You wa…

Other domains

D1 Agentic Architecture & Orchestration · D2 Tool Design & MCP Integration · D3 Claude Code Configuration & Workflows · D4 Prompt Engineering & Structured Output