CertSafari

    Free Practice Questions for Anthropic Claude Certified Developer – Professional (CCDV-P) Certification

    Exam guide version:
    1.0 (Effective November 2026)
    Guide checked for updates:
    20 Nov 2026
    Question bank created:
    19 Nov 2026
    Question bank last updated:
    20 Nov 2026

    Study with 441 exam-style practice questions designed to help you prepare for the Anthropic Claude Certified Developer – Professional (CCDV-P). All questions are aligned with the latest exam guide and include detailed explanations to help you master the material.

    Your progress

    Coverage
    Mastery
    Performance
    Lessons

    LessonsNew

    25 documentation-grounded lessons, one per exam-guide subdomain — every claim cited to the official docs.Based on the official docs as of 18 Nov 2026

    Start Practicing

    Start a quiz

    Practice with randomly mixed questions from all topics

    Question MixAll Topics
    FormatRandom Order

    Exam experiences

    Pass and fail outcomes from candidates who prepared here — advice, scores, and prep time.

    Your saved questions

    Open the list of questions you bookmarked during practice for this exam.

    Study notes

    Private notes per question, grouped by exam domain — opens on its own page, not inline on this overview.

    Quiz History

    Exam Details

    Key information about Anthropic Claude Certified Developer – Professional (CCDV-P)

    Question formats CertSafari offers
    • Multiple choice
    renewal:

    Free non-proctored renewal assessment on the Anthropic Partner Academy

    exam code:

    CCDV-P

    exam format:

    Scenario-based multiple-choice and multiple-response

    passing score:

    720 on a scale of 100–1,000

    prerequisites:

    None; Claude Certified Developer – Foundations recommended

    retake policy:

    Waits of 14, 30 and 90 days; up to 4 attempts per 12 months

    delivery method:

    Proctored: online (OnVUE) or Pearson test center

    target audience:

    Senior and staff software engineers, Platform and site reliability engineers, Solutions engineers

    registration fee:

    $175 USD

    result reporting:

    Pass/fail with scaled score and per-domain percent correct

    time limit minutes:

    120

    number of questions:

    70 (60 scored, 10 unscored)

    certification validity:

    12 months

    Exam Topics & Skills Assessed

    Skills measured (from the official study guide)

    1: Advanced Agent Architecture and Orchestration

    1.1: Multi-Agent Orchestration

    Evaluating orchestration patterns for production multi-agent designs and inter-agent failures, including supervisor/worker hierarchies, parallel subagent dispatch, delegation and context-isolation boundaries, inter-agent communication, and human-in-the-loop and approval flows.

    1.2: Durable Execution and State Management

    Applying session persistence, checkpointing, resumability, and memory architecture patterns to long-running agents that must survive interruption, restart, and model version change.

    1.3: Architecture Selection at Scale

    Evaluating production cost, reliability, and quality data to decide whether to evolve a running system between workflow and agentic patterns, and selecting the agent harness, framework, and reusable-pattern standards a team should adopt.

    1.4: Context and Memory Architecture

    Evaluating context-policy options for long-running agents approaching context limits or cost ceilings, including context compaction and editing, persistent memory, cache-safe instruction changes mid-conversation, and tool-use techniques that keep large tool catalogs out of the context window.

    2: Production Integration, Reliability, and Knowledge Grounding

    2.1: Reliability Engineering for Claude Services

    Evaluating and applying resilience patterns for production Claude integrations, including retry and backoff design, circuit breakers, graceful degradation, and cross-model, cross-region, and cross-platform fallbacks that distinguish rate limits from safety refusals from context-window stops while respecting per-platform feature parity.

    2.2: Scale and Throughput Management

    Analyzing rate-limit strategy, queueing and backpressure, batch pipeline design, multi-tenant quota and isolation, and capacity planning for high-volume Claude workloads, and identifying whether rate limits, token throughput, or batch economics is the binding constraint.

    2.3: Production MCP and Change Management

    Applying production MCP operation and change-management practice to MCP server deployments, model version migrations, and prompt updates, including remote MCP deployment and authentication, tool and schema versioning, migration testing against breaking behavior changes, platform-specific configuration parity and gateway/routing architectures, prompt versioning, and progressive rollout with rollback.

    2.4: Retrieval and Knowledge Grounding at Scale

    Evaluating retrieval architecture for knowledge-grounded production applications, including agentic search versus RAG tradeoffs, chunking and embedding strategy, hybrid search, and reranking, and designing grounding practices such as source citation, uncertainty declaration, and confidence signals, and diagnosing retrieval-quality failures by stage.

    3: Observability and Production Monitoring

    3.1: Telemetry and Instrumentation

    Applying instrumentation practice to production Claude systems under debuggability and privacy requirements, including distributed traces for agent loops and tool calls, token and cost metrics, cache-performance diagnostics, grounding and reasoning traceability, structured log events, and content-capture configuration under privacy constraints such as zero-data-retention.

    3.2: Production Monitoring and Alerting

    Analyzing production signals from LLM-backed services to define service level indicators (SLIs) and SLOs, design alerting and dashboards (including token and cost telemetry), and detect quality drift, behavioral regressions, and anomalous usage, including drift after model updates.

    3.3: Trace Analysis at Scale

    Analyzing high-volume trace data using sampling strategies, aggregation, and failure-cohort analysis to surface and quantify systemic issues that are not visible in individual traces.

    4: Advanced Debugging and Incident Response

    4.1: Complex Failure Diagnosis

    Analyzing evidence from complex production failures, including multi-turn, multi-agent, and non-deterministic failures, to isolate the problem origin among model output, agent harness, tools, and data, including model refusals as a failure class and retrieval-grounding failures.

    4.2: Incident Response and Learning Loops

    Evaluating incident-response options during an LLM-system outage or quality regression, including triage, mitigation, rollback, and communication, and converting the incident into regression evaluations, monitoring coverage, and postmortem findings.

    5: Evaluation Engineering and Quality Assurance

    5.1: Evaluation System Design

    Evaluating and designing offline evaluation systems for Claude applications, including golden sets, regression suites, behavioral checks, CI/CD evaluation gates, retrieval precision and groundedness metrics, and release criteria for prompt, model, and agent changes, and maintaining suite validity over time against contamination, saturation, and coverage gaps.

    5.2: LLM-as-Judge Engineering

    Analyzing LLM-as-judge evaluator quality against human-labeled data, including calibration, rubric design, bias and position effects, inter-rater reliability, and judge drift across model versions.

    5.3: Online Quality Measurement

    Analyzing quality on live production traffic using online evaluation techniques, including live-traffic sampling, A/B testing, feedback loops, and metric selection matched to task type (extraction, generation, agentic, retrieval-grounded).

    6: Cost and Latency Optimization at Scale

    6.1: Cost Architecture at Scale

    Evaluating fleet-level optimization strategies using usage and cost data, including cost modeling and forecasting, model routing and cascades across the Claude model family, caching strategy and cache diagnostics, batch-versus-realtime economics, and per-tenant budget enforcement.

    6.2: Latency Engineering

    Analyzing the latency profile of a Claude application against a latency target, distinguishing time-to-first-token from total latency, and applying streaming, parallel tool execution, cache checkpointing, and effort-level tuning.

    7: Runtime Guardrails and Production Security

    7.1: Runtime Guardrail Systems

    Evaluating layered guardrail architectures against a production agentʼs risk profile, including input/output filtering, PII redactors, hook-based policy enforcement, destructive-action prevention, sandboxing that grants tool and web access without exposing production systems, approval gates, refusal handling and safety fallbacks, deterministic enforcement for irreversible actions, and false-positive/false-negative tuning.

    7.2: Threat Defense and Adversarial Testing for Agentic Systems

    Analyzing the attack surface of tool-using agents that process untrusted content, including prompt injection, tool and memory poisoning, data exfiltration paths, unsafe-autonomy conditions, and the high-risk combination of private data access, untrusted input, and external communication, and applying defense-in-depth mitigations, red-team exercise design, and secure MCP configuration, drawing on the OWASP Top 10 for LLM Applications and the NIST AI Risk Management Framework and its Generative AI Profile.

    7.3: Identity, Secrets, and Audit at Scale

    Applying enterprise identity and secrets practice to production Claude deployments, including agent identity, workload identity federation replacing static API keys, credential and secrets management, least-privilege tool permissioning, and audit logging of security-relevant agent actions.

    8: Agent Fleet Operations for Software Delivery

    8.1: Parallel Agent Orchestration for Engineering Work

    Evaluating parallel fan-out patterns for large engineering tasks, including parallel agents on isolated worktrees, task decomposition, and merge and conflict strategy, and diagnosing fleet-level failure modes from fleet telemetry.

    8.2: Verification and AI Code Review Operations

    Analyzing verification operations in delivery pipelines where generation outpaces review, including AI code review introduced in shadow mode, calibration against human review before merge gating, and human oversight design.

    8.3: Unattended and Event-Driven Agent Operations

    Applying unattended-operation patterns to goal- or schedule-driven engineering tasks, including agents running toward goals, on schedules, or as event-driven responders such as CI/CD incident first response, with checkpointing, notification, and escalation design.

    8.4: Organization-Wide Coding Agent Configuration

    Applying enterprise configuration for organizations adopting coding agents, including managed settings and policies, MCP allowlists, telemetry, gateway routing, and governance with a quality and safety bar for sharing prompts, Skills, and MCP servers across teams.

    Techniques & products

    Claude API (Messages API)
    Claude Agent SDK
    Claude Managed Agents
    Claude Code
    Claude Code hooks
    Claude Code Skills
    Claude Code plugins
    Headless Claude Code
    Model Context Protocol (MCP)
    Remote MCP servers
    Claude model family
    Caching and cache diagnostics
    Batch processing
    Streaming
    Effort levels
    Context compaction
    Context editing
    Persistent memory
    Cloud-platform deployments
    Zero-data-retention
    Retrieval-augmented generation (RAG)
    Agentic search
    Hybrid search
    Reranking
    Embedding strategy
    Source citation
    Distributed tracing
    SLIs and SLOs
    LLM-as-judge
    Golden sets
    CI/CD evaluation gates
    A/B testing
    Model routing and cascades
    Circuit breakers
    Rate limiting and backpressure
    Multi-tenant quotas
    PII redaction
    Sandboxing
    Prompt injection
    Workload identity federation
    Secrets management
    Audit logging
    Isolated worktrees
    AI code review
    OWASP Top 10 for LLM Applications
    NIST AI Risk Management Framework
    NIST Generative AI Profile

    CertSafari is not affiliated with, endorsed by, or officially connected to Anthropic PBC. Full disclaimer