CCAR-P Study Resources
Study resources
The 132 documentation pages the CCAR-P lessons are built on, grouped by exam domain, with the lessons that cite each one. Then runnable notebooks from Anthropic's Claude Cookbook for hands-on practice.
Domain 1Solution Design & Architecture
- Define Success Criteria and Build EvaluationsLessons 1.1, 1.6
- Reduce HallucinationsLesson 1.1
- Tool Use OverviewLesson 1.1
- Building Effective AgentsLesson 1.1
- Building Effective AgentsLesson 1.2
- Use Case Guide: Ticket RoutingLesson 1.2
- Use Case Guide: Customer Support AgentLesson 1.2
- Demystifying Evals for AI AgentsLesson 1.2
- Building Effective Agents: The Augmented LLMLesson 1.3
- Building Effective Agents: WorkflowsLessons 1.3, 1.5
- Building Effective Agents: AgentsLesson 1.3
- Agent SDK OverviewLesson 1.3
- How We Built Our Multi-Agent Research SystemLesson 1.4
- Multi-Agent Research System: Production ReliabilityLesson 1.4
- Building Effective Agents: ParallelizationLesson 1.4
- Building Effective Agents: RoutingLesson 1.5
- Code Execution ToolLesson 1.5
- Agent SDK: Handle Approvals and User InputLesson 1.5
- Prompt Caching: PricingLesson 1.6
- Batch ProcessingLesson 1.6
- Reducing LatencyLesson 1.6
Domain 2Claude Models, Prompting & Context Engineering
- Choosing the Right ModelLesson 2.1
- Choosing the Right Model: Combine ModelsLesson 2.1
- Upgrade Between Model VersionsLesson 2.1
- Thinking: Steering and Cost ControlLesson 2.1
- Demystifying Evals for AI AgentsLesson 2.1
- Prompting Best Practices: Give Claude a RoleLesson 2.2
- Prompting Best Practices: Structure Prompts with XML TagsLesson 2.2
- Use Case Guide: Customer Support AgentLesson 2.2
- Structured OutputsLesson 2.2
- Mitigate Jailbreaks and Prompt InjectionsLesson 2.2
- Prompting Best Practices: Be Clear and DirectLesson 2.3
- Prompting Best Practices: Use Examples EffectivelyLesson 2.3
- Prompting Best Practices: Thinking and Step-by-Step ReasoningLesson 2.3
- Prompting Best Practices: Long-Context PromptingLesson 2.3
- Reduce HallucinationsLesson 2.3
- Context WindowsLesson 2.4
- Token CountingLesson 2.4
- Stop Reasons: Context Window ExceededLesson 2.4
- CompactionLesson 2.4
- Context EditingLesson 2.4
- Effective Context Engineering for AI AgentsLesson 2.4
- Prompt Caching: Structuring Your PromptLesson 2.5
- Prompt Caching: Pricing and Cache LifetimeLesson 2.5
- Prompt Caching: What Invalidates the CacheLesson 2.5
- Cache DiagnosticsLesson 2.5
- Agent Skills OverviewLesson 2.5
Domain 3Integration
- Tool Use OverviewLessons 3.1, 3.6
- Agent SDK Overview: Compare to Other Claude ToolsLesson 3.1
- Model Context Protocol: IntroductionLesson 3.1
- Model Context Protocol: ArchitectureLesson 3.1
- Claude on Google Cloud Vertex AILesson 3.1
- Securely Deploying AI Agents: Security BoundariesLesson 3.2
- Securely Deploying AI Agents: Least PrivilegeLesson 3.2
- Securely Deploying AI Agents: Credential ManagementLesson 3.2
- Model Context Protocol: AuthorizationLesson 3.2
- Mitigate Jailbreaks and Prompt Injections: Indirect InjectionLesson 3.2
- Writing Effective Tools for AI Agents: Choosing the Right ToolsLesson 3.3
- Writing Effective Tools for AI Agents: Tool DescriptionsLesson 3.3
- Writing Effective Tools for AI Agents: Token-Efficient ResponsesLesson 3.3
- Tool Search ToolLessons 3.3, 3.4
- Claude Code Subagents: Restricting ToolsLesson 3.3
- Effective Context Engineering: Just-in-Time ContextLessons 3.4, 3.6
- Agent Skills: Progressive DisclosureLesson 3.4
- Connect Claude Code to Tools via MCPLesson 3.4
- Contextual Retrieval: A Primer on RAGLesson 3.5
- Contextual Retrieval: Chunks That Lose Their ContextLesson 3.5
- Contextual Retrieval: When a Longer Prompt Is EnoughLesson 3.5
- EmbeddingsLesson 3.5
- CitationsLesson 3.5
- Contextual RetrievalLesson 3.6
- Contextual Retrieval: RerankingLesson 3.6
- Contextual Retrieval: Cost and LatencyLessons 3.6, 3.7
- Reducing LatencyLesson 3.7
- Thinking: Latency and Cost Trade-offsLesson 3.7
- Streaming MessagesLesson 3.7
- Parallel Tool UseLesson 3.7
- Batch ProcessingLesson 3.7
- Agent SDK: ObservabilityLesson 3.8
- Agent SDK: Control Sensitive Data in ExportsLesson 3.8
- Usage and Cost APILesson 3.8
- Workspaces: Per-Workspace Keys and LimitsLesson 3.8
- Rate LimitsLesson 3.8
Domain 4Evaluation, Testing & Optimization
- Define Success CriteriaLesson 4.1
- Define Success Criteria: Common CriteriaLesson 4.1
- Demystifying Evals for AI AgentsLesson 4.1
- Reducing Latency: How to Measure LatencyLessons 4.1, 4.5
- Build Evaluations: Design PrinciplesLesson 4.2
- Grade Your EvaluationsLesson 4.2
- Grade Your Evaluations: Tips for LLM-Based GradingLesson 4.2
- Demystifying Evals: Collecting TasksLesson 4.2
- Demystifying Evals: Types of GradersLesson 4.2
- Demystifying Evals: Evaluating on Live TrafficLessons 4.3, 4.5
- Demystifying Evals: Non-Determinism and Statistical ConfidenceLessons 4.3, 4.6
- Choosing a Model: Evaluate Before SwitchingLessons 4.3, 4.4
- Demystifying Evals: Capability and Regression EvalsLesson 4.3
- Reduce HallucinationsLesson 4.4
- Increase Output ConsistencyLesson 4.4
- Handling Stop ReasonsLesson 4.4
- Contextual Retrieval: Measuring Retrieval QualityLesson 4.4
- Agent SDK: Observability and AlertingLesson 4.5
- Model IDs and VersioningLesson 4.5
- Optimizing for Cost and IntelligenceLesson 4.6
- Reducing Latency: Optimize Prompt and Output LengthLesson 4.6
- Batch ProcessingLesson 4.6
Domain 5Governance, Safety & Risk Management
- Mitigate Jailbreaks and Prompt InjectionsLessons 5.1, 5.2
- Mitigate Jailbreaks: Indirect Prompt InjectionLesson 5.1
- Reduce HallucinationsLesson 5.1
- Reduce Prompt LeakLesson 5.1
- Securely Deploying AI AgentsLesson 5.1
- Mitigate Jailbreaks: Chaining SafeguardsLesson 5.2
- Mitigate Jailbreaks: Continuous MonitoringLesson 5.2
- Agent SDK: Permissions and Tool ApprovalLesson 5.2
- Claude Code SecurityLessons 5.2, 5.3
- Agent SDK: Handle Approvals and User InputLesson 5.3
- AI Fluency: Framework and FoundationsLessons 5.3, 5.5
- CitationsLesson 5.3
- API and Data Retention: PHI HandlingLesson 5.4
- API and Data Retention: Feature EligibilityLesson 5.4
- Data ResidencyLesson 5.4
- Anthropic Usage PolicyLessons 5.4, 5.5
- Anthropic Transparency HubLesson 5.4
- Evaluating and Mitigating Discrimination in Language Model DecisionsLesson 5.5
- Agent SDK: Logging and Tracing DecisionsLesson 5.5
Domain 6Stakeholder Communication & Lifecycle Management
- Define Success Criteria: Make Them MeasurableLessons 6.1, 6.3
- Define Success Criteria: Business OutcomesLessons 6.1, 6.5
- Building Effective Agents: Simplicity and TransparencyLesson 6.1
- Anthropic Trust CenterLesson 6.1
- Pricing: Forecasting Cost at VolumeLesson 6.2
- Reducing LatencyLesson 6.2
- Data ResidencyLessons 6.2, 6.3
- Building Effective AgentsLesson 6.2
- Demystifying Evals: Acting on Production SignalsLesson 6.3
- A Postmortem of Three Recent IssuesLesson 6.3
- Claude Code Memory: Recording Decisions and ConventionsLesson 6.4
- Claude in Amazon Bedrock: Regional EndpointsLesson 6.4
- Model DeprecationsLessons 6.4, 6.5
- Hosting the Agent SDK: Production ConcernsLesson 6.4
- Demystifying Evals: Capability and Regression EvalsLesson 6.5
- Agent SDK: Human Approval Before ActionsLesson 6.5
Domain 7Developer Productivity & Operational Enablement
- Claude Code Settings: Organization SettingsLesson 7.1
- Claude Code Memory: Where to Put CLAUDE.md FilesLesson 7.1
- Claude Code: Configure PermissionsLesson 7.1
- Manage Claude Code Plugins for Your OrganizationLesson 7.1
- Claude Code Champion Kit: Thirty-Day PlaybookLesson 7.1
- Claude Code: Track Team Usage with AnalyticsLesson 7.1
- Best Practices for Claude Code: Explore, Plan, CodeLesson 7.2
- Best Practices for Claude Code: Verify the WorkLesson 7.2
- Claude Code: Parallel Sessions with WorktreesLesson 7.2
- Run Claude Code ProgrammaticallyLesson 7.2
- Claude Code: Automate Actions with HooksLesson 7.2
- Agent SDK: Reading Agent TracesLesson 7.3
- Cache DiagnosticsLesson 7.3
- Usage and Cost APILesson 7.3
- Claude Code TroubleshootingLesson 7.3
Go hands-onClaude Cookbook notebooks
Optional and needs a Claude API key. Each notebook builds a pattern from the lessons listed beside it.
- Basic workflows: chaining, routing and parallelizationLesson 1.3
- Evaluator-optimizer workflowLesson 1.3
- Orchestrator-workers workflowLesson 1.4
- Metaprompt: draft a prompt template to start fromLesson 2.2
- Extracting structured JSONLesson 2.2
- Classification: from zero-shot to few-shot, with evaluationLesson 2.3
- Extended thinkingLesson 2.3
- Automatic context compactionLesson 2.4
- Context engineering toolsLesson 2.4
- Prompt cachingLesson 2.5
- Introduction to Agent SkillsLesson 2.5
- Customer service agent with tool useLesson 3.1
- Claude Agent SDK: a first agentLesson 3.1
- Tool search with embeddingsLessons 3.3, 3.4
- Controlling tool choiceLesson 3.3
- Memory and context managementLesson 3.4
- Retrieval-augmented generation, with evaluationLesson 3.5
- Contextual embeddingsLessons 3.5, 3.6
- CitationsLesson 3.6
- OpenTelemetry tracingLesson 3.8
- Usage and Cost APILesson 3.8
- Building evalsLessons 4.1, 4.2
- Generating test casesLesson 4.2
- Cost optimization for agentsLesson 4.6
- Batch processingLesson 4.6
- Building a moderation filterLesson 5.2
- Building custom SkillsLesson 7.1
- A site reliability agentLesson 7.3
Learning the topics? Start with the CCAR-P learning guide. Ready to test yourself? Take the practice tests.