Claude Certified Architect Foundations (CCAR-F) Practice Exam

These questions are part of the practice quiz. Start practicing

• Describe greenfield vs brownfield design considerations for architecture.
• Design patterns are repeatable solutions to common problems. Name two examples relevant to CCAR-F.
• What is the recommended approach to ensure traceability of claims with sources through summarization?
• Which technique maximizes first-pass success and reduces resubmission cost?
• What does a business capability map describe in relation to architecture?
• What is the role of an API contract in integration architecture?
• Which error category includes timeouts and service unavailability?
• Which API is described as offering 50% cost savings with up to a 24-hour processing window, but no guaranteed latency SLA and no multi-turn tool calling within a single request?
• To enable intelligent recovery decisions, what should a failure report include rather than a generic failure message?
• Which condition most clearly warrants escalation to a human reviewer rather than trusting automated results?
• Which artifact communicates the sequence of steps and milestones for realizing the target architecture?
• What mechanism exposes content catalogs to give agents visibility into available data without exploratory tool calls?
• Batch processing is appropriate for which type of workloads?
• Explain CAP theorem in the context of distributed systems and its relevance to CCAR-F decisions.
• Distinguish between synchronous vs asynchronous communication; provide CCAR-F context.
• Which frontmatter option restricts tool access during skill execution to prevent destructive actions?
• Which architecture features an orchestrator responsible for task decomposition, delegation, result aggregation, and dynamically selecting which subagents to invoke based on query complexity?
• Which term denotes emitting multiple Task tool calls in a single coordinator response to run subagents in parallel?
• What does 'Project-scoped MCP server' refer to?
• Which term denotes the mechanism for creating independent branches from a shared analysis baseline to explore divergent approaches (e.g., comparing two refactoring strategies)?
• What file defines each Agent Skill and includes YAML frontmatter for configuration?
• In distributed systems, what best describes log correlation and its importance?
• Which metrics are used to assess architectural quality attributes such as maintainability and modifiability?
• Which frontmatter option runs a skill in an isolated sub-agent context to prevent verbose outputs from polluting the main conversation?
• Which statement best describes path-scoped rules in Claude?
• Which term describes misrouting due to overlapping tool names?
• In CCAR-F, what is the purpose of Architecture Decision Records (ADRs)?
• Data governance and why is critical in CCAR-F?
• In an API-first approach with microservices, which practice helps ensure that changes to one service do not break others?
• Describe the 'two-pizza team' concept and its relevance to CCAR-F.
• What is maturity model in architecture governance?
• Which term refers to a subcomponent designed to isolate verbose discovery output and summarize results?
• Which term describes an orchestration model where a central component delegates tasks to subagents and aggregates results across iterations?
• Which practice supports avoiding contradictory findings across files?
• Which element is typically included in an Architectural Decision Record (ADR)?
• What is an Architectural Decision Record (ADR) and why is it important in CCAR-F?
• Where is the MCP server configuration stored?
• Which statement describes sentiment-based escalation pitfall?
• Which term describes the anti-pattern where a single coordinator manages all inter-subagent communication and routing between subagents?
• What security principle requires granting the minimum privileges necessary to perform a task?
• Which Claude Code CLI flag produces machine-parseable JSON output for automated processing such as inline PR comments?
• How are Agent Skills typically invoked in comparison to CLAUDE.md?
• Describe the concept of defense in depth in security design.
• In MCP, which property indicates whether a failed operation should be retried?
• What principle restricts each subagent's tool set to those relevant to its role?
• Which best describes attention dilution?
• What does test-driven iteration involve?
• In microservices data management, owning data stores leads to which consequence?
• Which mechanism allows optional fields to be omitted and return null when information is missing?
• Event storming aims to discover which elements in a system?
• What does ethics by design mean in architecture?
• What problem arises when one review pass analyzes many files at once?
• What is the term for a compiled summary provided when escalating to a human agent lacking the conversation transcript?
• What outcome does explicit linking of claims to their sources through summarization provide?
• What syntax organizes topic-specific rules as separate files instead of a single CLAUDE.md?
• Which set of error categories is used to guide agent responses in the MCP framework?
• When two customers share a name, what is the recommended action to disambiguate?
• Which approach guarantees schema-compliant structured output and eliminates JSON syntax errors?
• Which hook inspects outgoing tool calls to enforce compliance and redirect to escalation?
• What mechanism deterministically guarantees workflow ordering by leveraging hooks and prerequisite gates rather than prompt-based guidance?
• What is the primary objective of the CCAR-F Foundations exam?
• Describe disaster recovery planning (DRP) and its core components.
• In batch requests, what is the purpose of the custom_id field?
• Name three common architectural styles and briefly describe when each is appropriate.
• In the context of scalability strategies, what does stateless design imply?
• Which statement best describes the risk of using an LLM's self-reported confidence as an escalation trigger?
• Which objective is typically established to measure and manage expected service performance in CCAR-F using SRE practices?
• What triggers escalation in customer support context?
• Explain 'data privacy by design' and a CCAR-F example.
• What best describes the role of structured error context in recovery decisions?
• What are the three pillars of observability?
• What is threat modeling and which CCAR-F practice helps address it?
• If a system exhibits high false-positive rates in any one category, what is the effect on trust across categories?
• Using a specific flag in combination with --output-format json enforces a structured output schema in CI contexts.
• Stratified random sampling serves what purpose?
• Which prompting technique is described as the most effective for ensuring consistent outputs and generalizing to new patterns?
• Which term describes breaking a single customer message containing multiple issues into distinct items, investigating each in parallel using shared context, then synthesizing a unified resolution?
• Which mode forces the model to call some tool (but lets it choose which) and guarantees structured output when conversational text is unwanted?
• Which of the following is a common data anonymization technique used to protect privacy by removing direct identifiers?
• Having Claude ask questions to surface design considerations before implementing is called what?
• Which of the following best describes the role of SRE practices within CCAR-F foundations?
• Replacing a generic tool with focused subtools such as extract_data_points, summarize_content, and verify_claim_against_source demonstrates which architectural pattern?
• Which principle is essential to secure, resilient architecture and is often combined with least privilege?
• How do you perform risk assessment in an architectural design?
• What mechanism allows a model to signal its certainty to improve routing decisions?
• Describe risk-based security design approach.
• Which statement best captures a trade-off between monolithic and distributed deployments?
• Scratchpad content should address which issue in long explorations?
• Which statement describes test-driven iteration in terms of workflow?
• What is the purpose of the 'paths' field in YAML frontmatter for Claude rules?
• What phenomenon describes focusing on beginning and end while neglecting middle sections?
• In data architecture, which store is best for holding diverse, raw data for later processing and discovery?
• What is idempotency and why is it important in API design?
• Which term describes a multi-agent anti-pattern in which a single coordinator manages all inter-subagent communication, error handling, and information routing?
• What is 'change management' and why is it important in CCAR-F?
• What is event storming?
• Which statement best differentiates a service-oriented architecture (SOA) from microservices?
• Which statement describes the latency guarantees and multi-turn capabilities of the Message Batches API?
• Which term describes transforming or normalizing tool results before processing?
• Which term is the configuration object for a subagent including its description, system prompt, and tool restrictions?
• What is ADR adoption used for?
• In what scenario is a case facts block used?
• In architectural quality metrics, what does change failure rate measure?
• What is the observability vs monitoring distinction?
• In a scenario where an identifier yields multiple matches, which approach is recommended?
• Which term describes the hook that enforces compliance by inspecting outgoing tool calls and redirecting when necessary?
• What is the purpose of an architecture runway in the project roadmap?
• Which field records which code constructs triggered a finding, enabling analysis of false positive patterns?
• In MCP, which flag indicates an error condition to the agent?
• Which slash command lists the memory and context files currently loaded, helping diagnose session inconsistencies?
• Where are project-scoped custom commands stored within the repository?
• Which syntax is used to inject credentials into MCP server config without committing secrets to version control?
• Which term describes emitting multiple Task tool calls in a single coordinator response so subagents execute concurrently?
• Which statement best describes a non-functional requirement?
• Which term refers to the Agent SDK configuration that restricts which tools an agent can call and must include 'Task' for a coordinator to invoke subagents?
• Which pattern defines a structured, machine-readable way to report tool failures back to the agent?
• Data Lake vs Data Warehouse, and when would you choose each?
• What architecture uses per-file local analysis passes plus a separate cross-file integration pass?
• Which privacy-preserving technique adds statistical noise to protect individuals in aggregate queries?
• Which statement best describes Event-Driven Architecture in CCAR-F?
• Which term is associated with the practice of spreading work among multiple subagents by issuing several Task calls together in one coordinator response?
• What measures the risk of temporal data differences in outputs being misinterpreted as contradictions?
• What is the impact of high false-positive rates in any single category on overall developer trust?
• Which outcome indicates effectiveness of sample-then-batch refinement?
• What is vendor lock-in risk and how can architecture mitigate it?
• Which option describes Direct Execution Mode?
• Self-correction validation helps flag discrepancies by comparing which totals?
• Which term describes a fixed sequential decomposition pattern that splits work into ordered steps, such as per-file analysis followed by cross-file integration?
• What is 'contract testing' and why is it used in API-enabled architectures?
• Which term is the configuration object for a subagent including its description, system prompt, and tool restrictions?
• What best describes scalability and the CCAR-F suggested strategies?
• Which statement correctly describes Custom Commands in the CLAUDE environment?
• What is the default tool selection mode in which the model decides whether to call a tool or return text?
• Which artifact communicates the intended future architecture and its rationale to stakeholders?
• Which statement best explains the benefit of using tool_use with JSON schema?
• Which pattern describes a coordinator that evaluates synthesis output for gaps, re-delegates to search and analysis subagents with targeted queries, and re-invokes synthesis until coverage is sufficient?
• In MCP, which statement about retry handling is true for business rule violations?
• In distributed tracing, what is the purpose of a correlation ID in logs?
• What best describes the purpose of per-file local analysis passes followed by a cross-file integration pass?
• What is the benefit of event-driven architecture (EDA) in cloud-native systems?
• Which mechanism best supports decoupling between producers and consumers in integration design?
• Which term refers to an agent SDK hook that intercepts tool results to transform or normalize them before processing?
• Which pattern describes Claude reasoning about which tool to call next based on context, rather than following pre-configured decision trees?
• Which configuration requires a specific named tool to be called?
• What is a key benefit of keeping reasoning context in a model-generated code, paired with independent instances?
• Which term refers to renaming and rewriting ambiguous descriptions to fix misrouting?
• Field-level confidence calibration involves which practice?
• What is the primary goal of Layered Architecture in CCAR-F patterns?
• In CCAR-F, Enterprise Architecture is best described as a holistic view of an organization's strategy, processes, information, and technology, with artifacts including capability maps, architecture visions, roadmaps, and ADRs.
• Which statement about retry limits is correct?
• Which directory houses the Agent Skill definitions, each with a SKILL.md frontmatter?
• Which architectural style is characterized by asynchronous communication and decoupling?
• Which statement about confidence-tagged verification is true?
• What do path-scoped rules do?
• Scratchpad files are used to do what?
• Where is the MCP server configuration stored for personal or experimental use?
• Name the core non-functional requirements emphasized by CCAR-F.
• Which term corresponds to parsing natural language for termination signals and using arbitrary iteration caps as the primary stopping mechanism?
• Which fields are included in MCP error reporting to help retry, escalate, or explain appropriately?
• Which Claude mode is best for complex tasks such as large-scale changes across multiple files?
• What is the primary reason for calibrating review thresholds against a labeled validation set?
• Explain the difference between monolithic and distributed deployments with pros/cons.
• What is the purpose of trimming verbose tool output?
• What do RTO and RPO stand for in disaster recovery planning?
• Define refactoring.
• Explain the difference between functional requirements and non-functional requirements with an example.
• Explain SLOs and SLIs.
• How does the microservices architecture handle data management challenges?
• Which term addresses ambiguity or near-identical descriptions (e.g., analyze_content vs analyze_document) causing misrouting, by renaming and rewriting descriptions?
• What effect does providing too many tools to an agent have?
• Which mechanism keeps configuration modular across packages by referencing external files rather than a monolithic file?
• What is an Architecture View and its purpose in CCAR-F?
• In CCAR-F, how would you best describe an API gateway?
• What is idempotent design and provide an example in a distributed system?
• The lifecycle of sending a request to Claude, inspecting stop_reason, executing any requested tools, appending results to conversation history, and iterating until the model signals completion corresponds to which pattern?
• What is the primary purpose of including 2-4 targeted examples when prompting a model?
• Which feature handles environment variable expansion in the MCP configuration file .mcp.json?
• What is architecture runway?
• What is the role of a 'Capability Map' in CCAR-F?
• Which mode is most suitable for a single-file bug fix or a small validation change?
• Which concept is a fixed sequential decomposition pattern that breaks multi-aspect work into ordered steps?
• Self-correction validation can flag discrepancies by which approach?
• What is the purpose of modeling data lineage within an architectural design?
• What term best describes commands that are scoped to a specific project and accessed via slash syntax?
• Which statement best describes the purpose of stratified random sampling across document types and fields?
• Which term denotes the Claude Code CLI option that continues a specific named prior conversation?
• Which deployment approach reduces risk by progressively exposing the new version to a subset of users?
• Which term describes a programmatic block that prevents downstream tool calls until prerequisite steps complete?
• What is the difference between blue/green deployment and a canary release?
• Plan mode is appropriate for which type of task?
• Which flag enables JSON output suitable for machine parsing, often used with automation tools?
• What is the importance of the Cloud service model trio (IaaS, PaaS, SaaS) in CCAR-F?
• Which term describes task decomposition where subtasks are generated based on intermediate findings, suited to open-ended investigation tasks?
• Which option enforces a structured output schema in CI contexts when using JSON output?
• Which of the following is NOT among the listed non-functional requirements?