# AI Cert Guides: Full Content Listing

> Community study guides for Anthropic's Claude certifications: 123 lessons, 1,570 practice questions, mock exams, and build exercises across Claude Certified Architect (Foundations), Claude Certified Architect (Professional), Claude Certified Associate (Foundations), and Claude Certified Developer (Foundations).

URL: https://claudecertificationguide.com

---

## CCAR-F: Claude Certified Architect (Foundations)

Free study guide for the Claude Certified Architect (Foundations) exam, covering all 5 domains: Agentic Architecture, Tool Design & MCP, Claude Code, Prompt Engineering, and Context Management. 30 lessons and 407 practice questions.

### Domain 1: Agentic Architecture & Orchestration (27%)

Design and implement agentic systems using Claude's API, including loop management, orchestration patterns, guardrails, and the Claude Agent SDK.

#### 1.1 — Agentic Loops
Design and implement agentic loops for autonomous task execution using the Messages API, stop_reason field, and tool result handling.
URL: https://claudecertificationguide.com/learn/1-agentic-architecture/1-1-agentic-loops

#### 1.2 — Multi-Agent Orchestration
Orchestrate multi-agent systems with coordinator-subagent patterns, hub-and-spoke architecture, and proper isolation principles.
URL: https://claudecertificationguide.com/learn/1-agentic-architecture/1-2-orchestration-patterns

#### 1.3 — Subagent Invocation and Context Passing
Configure subagent invocation using the Task tool (renamed Agent in Claude Code v2.1.63), explicit context passing with structured metadata, parallel spawning, and fork_session.
URL: https://claudecertificationguide.com/learn/1-agentic-architecture/1-3-subagent-invocation-context

#### 1.4 — Workflow Enforcement and Handoff
Implement multi-step workflows with programmatic enforcement, prerequisite gates, and structured handoff protocols for human escalation.
URL: https://claudecertificationguide.com/learn/1-agentic-architecture/1-4-workflow-enforcement-handoff

#### 1.5 — Agent SDK Hooks
Apply Agent SDK hooks for data normalisation and deterministic policy enforcement using the PostToolUse and PreToolUse hook events.
URL: https://claudecertificationguide.com/learn/1-agentic-architecture/1-5-agent-sdk-hooks

#### 1.6 — Task Decomposition Strategies
Design task decomposition strategies using fixed sequential pipelines and dynamic adaptive decomposition, and solve the attention dilution problem.
URL: https://claudecertificationguide.com/learn/1-agentic-architecture/1-6-task-decomposition

#### 1.7 — Session State and Resumption
Manage session state using named session resumption, fork_session for divergent exploration, and fresh starts with summary injection to avoid stale context.
URL: https://claudecertificationguide.com/learn/1-agentic-architecture/1-7-session-state-resumption

### Domain 2: Tool Design & MCP Integration (18%)

Design effective tool schemas, implement MCP servers and clients, and integrate external services into Claude-powered applications.

#### 2.1 — Tool Interface Design
Designing effective tool interfaces with clear descriptions and boundaries for reliable LLM tool selection
URL: https://claudecertificationguide.com/learn/2-tool-design-mcp/2-1-tool-schema-design

#### 2.2 — Structured Error Responses
Implementing structured error responses for MCP tools with proper categorisation and recovery metadata
URL: https://claudecertificationguide.com/learn/2-tool-design-mcp/2-2-structured-error-responses

#### 2.3 — Tool Distribution & Tool Choice
Distributing tools across agents and configuring tool_choice for reliable tool selection
URL: https://claudecertificationguide.com/learn/2-tool-design-mcp/2-3-tool-distribution-choice

#### 2.4 — MCP Server Integration
Integrating MCP servers into Claude Code and agent workflows with proper scoping and configuration
URL: https://claudecertificationguide.com/learn/2-tool-design-mcp/2-4-mcp-server-integration

#### 2.5 — Built-in Tools
Selecting and applying built-in tools (Read, Write, Edit, Bash, Grep, Glob) effectively for codebase operations
URL: https://claudecertificationguide.com/learn/2-tool-design-mcp/2-5-built-in-tools

### Domain 3: Claude Code Configuration & Workflows (20%)

Configure Claude Code for development workflows, manage settings, hooks, permissions, and integrate with CI/CD pipelines.

#### 3.1 — CLAUDE.md Hierarchy, Scoping, and Modular Organisation
Master the three-level CLAUDE.md configuration hierarchy, understand scoping rules, and learn modular organisation patterns with @path imports and .claude/rules/
URL: https://claudecertificationguide.com/learn/3-claude-code-config/3-1-claude-md-hierarchy

#### 3.2 — Custom Slash Commands and Skills
Learn the unified Skills system for custom commands, configure SKILL.md frontmatter, and understand when to use skills vs CLAUDE.md
URL: https://claudecertificationguide.com/learn/3-claude-code-config/3-2-slash-commands-skills

#### 3.3 — Path-Specific Rules for Conditional Convention Loading
Master path-specific rules in .claude/rules/ with YAML frontmatter glob patterns for conditional convention loading, token efficiency, and cross-directory coverage
URL: https://claudecertificationguide.com/learn/3-claude-code-config/3-3-path-specific-rules

#### 3.4 — Plan Mode vs Direct Execution
Learn the decision framework for choosing plan mode or direct execution, understand the Explore subagent, and master the hybrid plan-then-execute pattern
URL: https://claudecertificationguide.com/learn/3-claude-code-config/3-4-plan-mode-execution

#### 3.5 — Iterative Refinement Techniques
Master the technique hierarchy for iterative refinement: concrete examples over prose, test-driven iteration, the interview pattern, and batch vs sequential feedback
URL: https://claudecertificationguide.com/learn/3-claude-code-config/3-5-iterative-refinement

#### 3.6 — CI/CD Integration
Master running Claude Code in CI/CD pipelines: the -p flag for non-interactive mode, structured JSON output, session isolation for reviews, and incremental review context
URL: https://claudecertificationguide.com/learn/3-claude-code-config/3-6-cicd-integration

### Domain 4: Prompt Engineering & Structured Output (20%)

Craft effective prompts, implement structured output patterns, and apply prompt engineering techniques for production Claude applications.

#### 4.1 — System Prompts with Explicit Criteria
Design prompts with explicit criteria to improve precision and reduce false positives
URL: https://claudecertificationguide.com/learn/4-prompt-engineering/4-1-system-prompts

#### 4.2 — Few-Shot Prompting
Apply few-shot prompting to improve output consistency and quality
URL: https://claudecertificationguide.com/learn/4-prompt-engineering/4-2-few-shot-prompting

#### 4.3 — Structured Output with Tool Use
Enforce structured output using tool use and JSON schemas
URL: https://claudecertificationguide.com/learn/4-prompt-engineering/4-3-structured-output

#### 4.4 — Validation, Retry, and Feedback Loops
Implement validation, retry, and feedback loops for extraction quality
URL: https://claudecertificationguide.com/learn/4-prompt-engineering/4-4-validation-retry-loops

#### 4.5 — Batch Processing Strategies
Design efficient batch processing strategies
URL: https://claudecertificationguide.com/learn/4-prompt-engineering/4-5-batch-processing

#### 4.6 — Multi-Instance and Multi-Pass Review
Design multi-instance and multi-pass review architectures
URL: https://claudecertificationguide.com/learn/4-prompt-engineering/4-6-multi-pass-review

### Domain 5: Context Management & Reliability (15%)

Manage context windows effectively, implement caching strategies, handle long conversations, and build reliable production systems.

#### 5.1 — Context Window Management
Manage conversation context to preserve critical information across long interactions, including the progressive summarisation trap, lost-in-the-middle effect, and tool result trimming.
URL: https://claudecertificationguide.com/learn/5-context-management/5-1-context-window-management

#### 5.2 — Escalation & Ambiguity Resolution
Design effective escalation and ambiguity resolution patterns, including valid escalation triggers, unreliable triggers, frustration nuance, and ambiguous customer matching.
URL: https://claudecertificationguide.com/learn/5-context-management/5-2-escalation-ambiguity

#### 5.3 — Error Propagation in Multi-Agent Systems
Implement error propagation strategies across multi-agent systems, including structured error context, the distinction between access failures and valid empty results, and coverage annotations.
URL: https://claudecertificationguide.com/learn/5-context-management/5-3-error-propagation

#### 5.4 — Codebase Exploration & Context Degradation
Manage context effectively in large codebase exploration, including context degradation mitigation, scratchpad files, subagent delegation, and crash recovery via structured state manifests.
URL: https://claudecertificationguide.com/learn/5-context-management/5-4-codebase-exploration

#### 5.5 — Human Review & Confidence Calibration
Design human review workflows and confidence calibration, including the aggregate metrics trap, stratified random sampling, field-level confidence scores, and reviewer capacity prioritisation.
URL: https://claudecertificationguide.com/learn/5-context-management/5-5-human-review-calibration

#### 5.6 — Information Provenance & Multi-Source Synthesis
Preserve information provenance and handle uncertainty in multi-source synthesis, including structured claim-source mappings, conflict handling, temporal awareness, and content-appropriate rendering.
URL: https://claudecertificationguide.com/learn/5-context-management/5-6-information-provenance

---

## CCAR-P: Claude Certified Architect (Professional)

Free study guide for the Claude Certified Architect (Professional) exam, covering all 7 domains: Solution Design, Models & Prompting, Integration, Evaluation & Testing, Governance & Safety, Stakeholders & Lifecycle, and Developer Productivity. 38 lessons and 397 practice questions.

### Domain 1: Solution Design & Architecture (17%)

Translate business problems into end-to-end Claude architectures: architectural patterns, multi-agent design, decomposition, and business-value alignment.

#### 1.1 — Translating Business Problems into Claude Solutions
Turn business problems into well-scoped Claude solutions: capture constraints, define success criteria, map requirements to task shapes, and draw the boundary between deterministic code and the model.
URL: https://claudecertificationguide.com/ccar-p/learn/1-solution-design/1-1-translating-business-problems

#### 1.2 — End-to-End Architectures and Feedback Loops
Design the six-stage production pipeline — input, context assembly, processing, output validation, delivery, and feedback loops — and treat model output as untrusted until validated.
URL: https://claudecertificationguide.com/ccar-p/learn/1-solution-design/1-2-end-to-end-architectures

#### 1.3 — Selecting Architectural Patterns
Choose between augmented LLM, workflow, and agentic architectures using Anthropic's canonical taxonomy — and apply the simplest-solution decision rule the exam rewards.
URL: https://claudecertificationguide.com/ccar-p/learn/1-solution-design/1-3-architectural-patterns

#### 1.4 — Multi-Agent Systems and Orchestration
Design orchestrator-worker systems that earn their token multiplier: explicit delegation, context isolation, effort scaling, and root-cause diagnosis of multi-agent failures.
URL: https://claudecertificationguide.com/ccar-p/learn/1-solution-design/1-4-multi-agent-systems

#### 1.5 — Decomposition Techniques for Complex Problems
Split complex problems along the right axis, fix attention dilution with local-plus-integration passes, and keep stage interfaces structured — while knowing when not to decompose at all.
URL: https://claudecertificationguide.com/ccar-p/learn/1-solution-design/1-5-decomposition-techniques

#### 1.6 — Aligning Solutions to Business Value Pillars
Tie every architecture decision to a business value pillar — efficiency, transformation, productivity, cost, performance SLAs — and engineer SLAs and unit economics rather than hoping for them.
URL: https://claudecertificationguide.com/ccar-p/learn/1-solution-design/1-6-business-value-pillars

### Domain 2: Claude Models, Prompting & Context Engineering (13%)

Select Claude models on their trade-offs, and engineer the system prompts, guardrails, and context strategies that guide them.

#### 2.1 — Selecting Claude Models on Trade-Offs
Choose models as constraint satisfaction, not prestige: right-size per pipeline stage with eval evidence, design escalation cascades, and treat model lifecycle as an architecture concern.
URL: https://claudecertificationguide.com/ccar-p/learn/2-models-prompting/2-1-model-selection-trade-offs

#### 2.2 — System Prompts, Templates, and Guardrails
Treat the system prompt as an architectural contract: explicit role and scope, an escape hatch against hallucination, XML-structured templates, and guardrails layered in code around the model.
URL: https://claudecertificationguide.com/ccar-p/learn/2-models-prompting/2-2-system-prompts-and-guardrails

#### 2.3 — Zero-Shot, Few-Shot, and Chain-of-Thought
Climb the technique ladder in cost order — instructions, examples, reasoning — and match each technique to the failure it actually fixes, knowing where built-in thinking absorbs prompted CoT.
URL: https://claudecertificationguide.com/ccar-p/learn/2-models-prompting/2-3-prompt-engineering-techniques

#### 2.4 — Optimising Context Windows and Token Usage
Budget the window like the finite resource it is: count what occupies it, fight context rot with curation, attack the replay bill, and know why caching never buys window space.
URL: https://claudecertificationguide.com/ccar-p/learn/2-models-prompting/2-4-context-window-optimisation

#### 2.5 — Prompt Reuse: Caching, Modular Prompts, and Skills
Front-load stability so prefixes cache, run the break-even arithmetic on TTLs, share prompt sections as versioned modules, and load deep expertise on demand through Skills.
URL: https://claudecertificationguide.com/ccar-p/learn/2-models-prompting/2-5-prompt-reuse-strategies

### Domain 3: Integration (19%)

Integrate Claude into enterprise systems: security analysis, observability at scale, RAG pipeline design, and choosing the right connection protocol.

#### 3.1 — Capability Bloat and Least-Privilege Configuration
Evaluate agent tool surfaces with least privilege: every tool the role never needs is attack surface plus context weight, and removal beats every mitigation the exam plants.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-1-capability-bloat

#### 3.2 — Authentication, Authorisation, and the Gaps Between
Trace identity hop by hop to find the real security gap: confused-deputy service accounts, secrets in model-visible context, and the token passthrough the MCP specification forbids.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-2-authentication-authorisation-gaps

#### 3.3 — Accuracy–Latency Trade-offs
Decompose the latency budget, price every accuracy lever, and justify configurations the professional way: both the SLA and the quality bar named, neither sacrificed to the other.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-3-accuracy-latency-trade-offs

#### 3.4 — Observability and Monitoring at Scale
Non-determinism breaks anecdote debugging: make the trace the unit of observability, sample payloads at scale, and turn answer quality into a monitored signal rather than a support-ticket surprise.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-4-observability-at-scale

#### 3.5 — RAG Pipeline Design: Chunking and Indexing
Design retrieval pipelines stage by stage: structure-aware chunking, contextual retrieval, hybrid indexing, and the operational freshness discipline that stops confident stale answers.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-5-rag-pipeline-design

#### 3.6 — Matching Retrieval to Data Shape and Query Pattern
Choose retrieval per source and per question: system-of-record tools for live data, semantic search for prose, agentic retrieval for multi-hop work — the cheapest strategy whose failure mode is acceptable.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-6-retrieval-strategy-matching

#### 3.7 — Choosing the Integration Mechanism
MCP, direct API, headless runs, agent-to-agent: count the consumers, test whether the counterpart reasons, and match the mechanism to the pattern instead of the prestige.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-7-integration-mechanisms

#### 3.8 — Progressive Discovery vs Monolithic Context
Decide what loads upfront and what loads on demand using four cues — surface-to-slice ratio, frequency, latency, stability — and the hybrid that keeps a hot cached core with the long tail behind discovery.
URL: https://claudecertificationguide.com/ccar-p/learn/3-integration/3-8-progressive-discovery

### Domain 4: Evaluation, Testing & Optimisation (16%)

Define evaluation metrics, build datasets and test frameworks, run A/B tests, and diagnose and optimise live Claude systems.

#### 4.1 — Defining Evaluation Metrics
Write the metric sheet before the system: five metric families with definitions, measurement methods, and baselines; success versus guard metrics; and the proxy trap that quietly rewrites your targets.
URL: https://claudecertificationguide.com/ccar-p/learn/4-evaluation-testing/4-1-defining-evaluation-metrics

#### 4.2 — Designing Evaluation Datasets and Test Frameworks
Compose datasets as risk coverage: four case populations, golden-set discipline, statistically honest sizing, and the pairing rule that assigns code-graded checks, LLM judges, and human review to the work each does best.
URL: https://claudecertificationguide.com/ccar-p/learn/4-evaluation-testing/4-2-evaluation-datasets-test-frameworks

#### 4.3 — A/B Testing and Iterative Improvement
Offline evals gate, online experiments validate: change one variable at a time, carry guard metrics on every experiment, meet the statistical bar before declaring wins, and stage exposure for risky swaps.
URL: https://claudecertificationguide.com/ccar-p/learn/4-evaluation-testing/4-3-ab-testing-iterative-improvement

#### 4.4 — Diagnosing Prompt Failure, Hallucination, and Model Mismatch
Run the diagnostic order the exam rewards: trace and reproduce, isolate the stage the evidence implicates, classify the failure signature, run the cheapest decisive test, fix the failing layer, and mint regression cases.
URL: https://claudecertificationguide.com/ccar-p/learn/4-evaluation-testing/4-4-diagnosing-system-issues

#### 4.5 — Optimising Token Usage, Latency, and Cost-Performance Trade-offs
Measure before optimising and re-gate after: match the lever to the dominant term — caching, curation, effort budgets, streaming, right-sizing, batch — and always name what the optimisation spends to save what.
URL: https://claudecertificationguide.com/ccar-p/learn/4-evaluation-testing/4-5-token-latency-cost-optimisation

#### 4.6 — Monitoring System Performance in Production
Operate the quality practice on Domain 3's plumbing: continuous evaluation of sampled traffic, drift indicators that move before complaints, alerts on guard metrics at the tail, and the incident loop that mints regression cases.
URL: https://claudecertificationguide.com/ccar-p/learn/4-evaluation-testing/4-6-monitoring-observability

### Domain 5: Governance, Safety & Risk Management (14%)

Guardrails and safety controls, risk and failure-mode identification, human-in-the-loop validation, regulatory compliance, and ethical AI.

#### 5.1 — Guardrails, Safety Controls, and the Surface a Risk Belongs To
Assign a named risk to the control surface that can actually hold it: input screening, the instruction layer, capability scoping, output validation, or not building the capability at all.
URL: https://claudecertificationguide.com/ccar-p/learn/5-governance-safety/5-1-guardrails-and-safety-controls

#### 5.2 — Risks, Limitations, and the Failure Modes That Stay Silent
Turn a list of things that could go wrong into a risk register that changes the design: mechanism, blast radius, detection signal, owning control, and the person who accepted the residual.
URL: https://claudecertificationguide.com/ccar-p/learn/5-governance-safety/5-2-risks-limitations-failure-modes

#### 5.3 — Human-in-the-Loop Validation: Placement, Pattern, and Proof
Place the human gate where the action cannot be undone, pick between the five oversight patterns on reversibility and blast radius, and prove the gate is real by measuring what it overrides.
URL: https://claudecertificationguide.com/ccar-p/learn/5-governance-safety/5-3-human-in-the-loop-validation

#### 5.4 — Regulatory Compliance as an Architecture Constraint
Translate an obligation into a design constraint: draw scope around the whole data path, minimise at the boundary, get the agreement in place before the data moves, and know whose authorisation boundary you are inside.
URL: https://claudecertificationguide.com/ccar-p/learn/5-governance-safety/5-4-regulatory-compliance

#### 5.5 — Ethical AI: Bias, Fairness, and Transparency
Treat bias as a measured property of outcomes, choose and state a fairness criterion, work the remediation ladder in order, and deliver each of the three transparency obligations where it is actually owed.
URL: https://claudecertificationguide.com/ccar-p/learn/5-governance-safety/5-5-ethical-ai-bias-fairness-transparency

### Domain 6: Stakeholder Communication & Lifecycle Management (14%)

Structured discovery, communicating architectural decisions and trade-offs, and supporting solutions through every lifecycle phase.

#### 6.1 — Structured Discovery and Requirement Gathering
Find the problem behind the request: the measured baseline, the consequence of a wrong output, the data and lawful-basis picture, and criteria that can be passed or failed.
URL: https://claudecertificationguide.com/ccar-p/learn/6-stakeholder-lifecycle/6-1-structured-discovery

#### 6.2 — Communicating Architectural Decisions and Trade-offs
State the decision, the options, the criteria, the evidence and what would change your mind — then leave the choice where the accountability sits, unless an obligation removes it.
URL: https://claudecertificationguide.com/ccar-p/learn/6-stakeholder-lifecycle/6-2-communicating-decisions-trade-offs

#### 6.3 — Expectation Alignment and Service-Level Commitments
Measure the envelope before you promise anything, write commitments that carry all six elements, and run a feedback loop that produces evidence rather than sentiment.
URL: https://claudecertificationguide.com/ccar-p/learn/6-stakeholder-lifecycle/6-3-expectation-alignment-and-slas

#### 6.4 — Documentation and Implementation Guidance
Hand over the judgement rather than the history: the artefact set that lets a stranger change a system safely, and the baseline that tells them when they have broken it.
URL: https://claudecertificationguide.com/ccar-p/learn/6-stakeholder-lifecycle/6-4-documentation-and-implementation-guidance

#### 6.5 — Supporting the Lifecycle Phases
Gate each phase on what it produces rather than when it is due, transfer the capability to own rather than the code, and treat the events you can see coming as planned re-validation.
URL: https://claudecertificationguide.com/ccar-p/learn/6-stakeholder-lifecycle/6-5-lifecycle-phases

### Domain 7: Developer Productivity & Operational Enablement (7%)

Equip teams with Claude tooling and AI-assisted workflows, and support debugging and operational issue resolution.

#### 7.1 — Team Tooling Configuration: Guidance Versus Enforcement
Sort every element of a team's Claude tooling setup into the layer that fits it, and know why a written instruction cannot prevent an operation that an enforced rule can.
URL: https://claudecertificationguide.com/ccar-p/learn/7-developer-productivity/7-1-team-tooling-configuration

#### 7.2 — AI-Assisted Developer Workflows: Measure the Work, Package the Practice
Judge AI-assisted development on outcome measures rather than adoption figures, automate the work that has a cheap check, and package proven practice so quality stops depending on who ran it.
URL: https://claudecertificationguide.com/ccar-p/learn/7-developer-productivity/7-2-ai-assisted-developer-workflows

#### 7.3 — Debugging and Operational Resolution: Evidence Before Change
Lead a response to a misbehaving production agent with evidence rather than instinct, decide containment on blast radius, and leave the incident with an artefact the team inherits.
URL: https://claudecertificationguide.com/ccar-p/learn/7-developer-productivity/7-3-debugging-and-operational-resolution

---

## CCAO-F: Claude Certified Associate (Foundations)

Free study guide for the Claude Certified Associate (Foundations) exam, covering all 7 domains: Prompting, Output Evaluation, Products & Models, Workflow Integration, Configuration, Governance & Risk, and Troubleshooting. 30 lessons and 397 practice questions.

### Domain 1: Prompting and Task Execution (14%)

Create effective prompts for business and technical tasks, decompose complex requests, and iterate on outputs across task types.

#### 1.1 — Creating Effective Prompts for Business and Technical Tasks
Brief Claude the way you would brief a capable new colleague: state the task, supply context and source material, specify the output, and add an example when standards matter.
URL: https://claudecertificationguide.com/ccao-f/learn/1-prompting-task-execution/1-1-effective-prompts

#### 1.2 — Decomposing Complex Requests
Structure complex, multi-part requests with numbered steps, staged conversations, plan-first prompting, and separate chats so Claude can do each piece properly.
URL: https://claudecertificationguide.com/ccao-f/learn/1-prompting-task-execution/1-2-task-decomposition

#### 1.3 — Iterating Prompts to Improve Output
Treat first outputs as drafts and improve them with specific feedback: name the gap, ask for variants, and know when to refine and when to re-brief.
URL: https://claudecertificationguide.com/ccao-f/learn/1-prompting-task-execution/1-3-prompt-iteration

#### 1.4 — Adapting Strategy to Task Type
Match the briefing shape to the work: material and questions for analysis, scoped questions for research, full briefs for drafting, and wide-then-narrow prompts for brainstorming.
URL: https://claudecertificationguide.com/ccao-f/learn/1-prompting-task-execution/1-4-task-type-strategies

### Domain 2: Output Evaluation and Validation (21%)

The exam's largest domain: evaluate Claude's output for accuracy, spot hallucinations and bias, and know when human review is required.

#### 2.1 — Evaluating Outputs for Accuracy and Completeness
Treat every Claude output as a strong first draft from a colleague who can be confidently wrong: check the specifics against the source, then walk the source's own structure for anything missing.
URL: https://claudecertificationguide.com/ccao-f/learn/2-output-evaluation/2-1-accuracy-completeness

#### 2.2 — Identifying Hallucinations, Inconsistencies and Bias
Learn the three failure types by shape: hallucination is fluent and fabricated, inconsistency is the output disagreeing with itself, and bias is a skewed pattern across descriptions of people.
URL: https://claudecertificationguide.com/ccao-f/learn/2-output-evaluation/2-2-hallucinations-inconsistencies-bias

#### 2.3 — Fact-Checking and Validation Techniques
Build the validation toolkit from cheapest to heaviest: check against the source you gave Claude, verify external claims against primary documents, make outputs traceable, and match the effort to the stakes.
URL: https://claudecertificationguide.com/ccao-f/learn/2-output-evaluation/2-3-fact-checking-validation

#### 2.4 — When Human Review or Additional Verification Is Required
Triage outputs by what they carry, not how they read: escalate to a reviewer qualified in the subject when the stakes demand it, and let your own careful review suffice when they do not.
URL: https://claudecertificationguide.com/ccao-f/learn/2-output-evaluation/2-4-human-review-escalation

#### 2.5 — Editing, Adapting and Comparing Outputs for the Audience
Own the editorial verdict: recognise when an accurate output is wrong for its reader, adapt it without losing the substance, and compare versions against criteria fixed before you read them.
URL: https://claudecertificationguide.com/ccao-f/learn/2-output-evaluation/2-5-editing-adapting-outputs

#### 2.6 — Curating Information and Selecting Output Formats
Turn working material into a work product: extract what the deliverable needs from the conversation, then match the format to the reader's use, inline for the moment, a table for comparison, an artifact for anything with a life of its own.
URL: https://claudecertificationguide.com/ccao-f/learn/2-output-evaluation/2-6-output-formats-curation

### Domain 3: Product and Model Selection (12%)

Choose the right Claude product feature and model for the task, and manage context limitations and memory.

#### 3.1 — Selecting the Right Claude Product Feature
Read the shape of the task before reaching for a tool: ordinary chat for one-offs, Projects for recurring context, Research for cited multi-source questions, and artifacts for work with its own life.
URL: https://claudecertificationguide.com/ccao-f/learn/3-product-model-selection/3-1-product-feature-selection

#### 3.2 — Differentiating Between Claude Model Types
Haiku-class, Sonnet-class and Opus-class models span a real trade-off between speed, cost and depth; the tested skill is reading which end of that trade-off a task actually needs.
URL: https://claudecertificationguide.com/ccao-f/learn/3-product-model-selection/3-2-model-tiers

#### 3.3 — Aligning Model Choice with Cost, Speed and Quality
Every model choice spends three currencies, cost, time and depth; name the one the task actually needs, pay only that, and split mixed workloads by their demands.
URL: https://claudecertificationguide.com/ccao-f/learn/3-product-model-selection/3-3-cost-speed-quality

#### 3.4 — Managing Context Limits and Memory: Restart, Summarise, Persist
A conversation holds a finite working space; plan its end the way you plan a long meeting: restart cleanly, summarise and verify the handover, and persist the outputs the next chat needs.
URL: https://claudecertificationguide.com/ccao-f/learn/3-product-model-selection/3-4-context-and-memory-management

### Domain 4: Workflow Integration and Solution Design (16%)

Bring Claude into real workflows: requirements analysis, research and planning, solution design, and communicating value to stakeholders.

#### 4.1 — Analysing Requirements and Use Cases
Use Claude to turn messy input into structured requirements, then judge which pieces of work are genuinely Claude-shaped: Claude structures, the human validates and owns the result.
URL: https://claudecertificationguide.com/ccao-f/learn/4-workflow-integration/4-1-requirements-and-use-cases

#### 4.2 — Using Claude for Research, Planning and Process Optimisation
Three applications, one shape: Claude produces the first pass and you own the judgement. State the real constraints, and change nothing until the evidence names the bottleneck.
URL: https://claudecertificationguide.com/ccao-f/learn/4-workflow-integration/4-2-research-planning-optimisation

#### 4.3 — Supporting Solution Design, Development and Iteration
Iterate the deliverable, not the prompt: run each round against one named criterion drawn from the requirement, compare concrete versions, and stop when the remaining changes are preference.
URL: https://claudecertificationguide.com/ccao-f/learn/4-workflow-integration/4-3-solution-design-iteration

#### 4.4 — Integrating Claude into Existing Workflows
Decide where in an existing process a person brings work to Claude: augment one step or redesign the process, keep a human on the deciding step, and escalate what needs systems or unattended running.
URL: https://claudecertificationguide.com/ccao-f/learn/4-workflow-integration/4-4-integrating-existing-workflows

#### 4.5 — Communicating Value and Limitations to Stakeholders
Say what is true, including the parts that are not impressive: measured claims, the untested category named, caught errors presented as evidence the process works, and each stakeholder given what they need.
URL: https://claudecertificationguide.com/ccao-f/learn/4-workflow-integration/4-5-communicating-value-limitations

### Domain 5: Configuration and Knowledge Management (12%)

Configure Claude Projects with instructions, knowledge sources, and connectors, and keep those configurations current.

#### 5.1 — Configuring Projects with Instructions and Knowledge Sources
A Project has two containers doing different jobs: custom instructions govern how Claude behaves in every chat, and project knowledge supplies the material it works from. Sort every piece by which job it does.
URL: https://claudecertificationguide.com/ccao-f/learn/5-configuration-knowledge/5-1-project-instructions-knowledge

#### 5.2 — Managing Uploaded Knowledge and Connectors
Material reaches a Project by two routes: an uploaded copy that is stable but goes stale silently, or a connector that reads the live source. Choose by how the material changes, and curate rather than dump.
URL: https://claudecertificationguide.com/ccao-f/learn/5-configuration-knowledge/5-2-uploads-and-connectors

#### 5.3 — Creating Effective System-Level Instructions
Standing instructions live at three levels: your account, a Project, and the organisation. Place each rule at the level it is true of, and write it as a behaviour someone could check.
URL: https://claudecertificationguide.com/ccao-f/learn/5-configuration-knowledge/5-3-system-level-instructions

#### 5.4 — Informing, Maintaining, and Updating Configurations
A configuration is a living thing: knowledge goes stale, instructions outlive the world they describe, and scope creeps. Remove superseded material, sweep the instructions, and tell the people who use it.
URL: https://claudecertificationguide.com/ccao-f/learn/5-configuration-knowledge/5-4-maintaining-configurations

### Domain 6: Governance, Risk, and Responsible Use (15%)

Use Claude appropriately: data sensitivity, privacy and regulatory considerations, organisational AI policy, and ethics.

#### 6.1 — Identifying Appropriate and Inappropriate Use Cases
Fit asks whether Claude would do the work well; appropriateness asks whether anyone should hand it over at all. Learn the four inappropriate categories, the substitution test, and why over-caution is its own failure.
URL: https://claudecertificationguide.com/ccao-f/learn/6-governance-responsible-use/6-1-appropriate-use-cases

#### 6.2 — Applying Data Sensitivity, Regulatory and Privacy Considerations
Recognise sensitive material, then climb the ladder: remove it, anonymise it, use an authorised route, or ask the owner. The safeguard happens before the upload; nothing typed into the chat governs it afterwards.
URL: https://claudecertificationguide.com/ccao-f/learn/6-governance-responsible-use/6-2-data-sensitivity-privacy

#### 6.3 — Following Organisational AI Policies and Governance Standards
The sanctioned workspace is substance, not etiquette: the organisation's agreement, administration and controls attach to it and reach no personal account. Learn the two-half shadow-AI response and why policy silence is not permission.
URL: https://claudecertificationguide.com/ccao-f/learn/6-governance-responsible-use/6-3-organisational-ai-policies

#### 6.4 — Understanding the Ethical Implications of AI Usage
The name on the work owns it completely; saying Claude drafted it transfers nothing. Learn when disclosure is genuinely owed, why fairness is a different review from accuracy, and why colleagues are stakeholders too.
URL: https://claudecertificationguide.com/ccao-f/learn/6-governance-responsible-use/6-4-ethical-implications

### Domain 7: Troubleshooting and Optimisation (10%)

Diagnose underperforming prompts and poor outputs, adjust approach based on feedback, and optimise workflows.

#### 7.1 — Diagnosing and Resolving Underperforming Prompts and Poor Outputs
Ask the routing question before touching anything: a one-off disappointment points at the request, a defect in every output points at the setup, and each cause on the short list has its own fix.
URL: https://claudecertificationguide.com/ccao-f/learn/7-troubleshooting-optimisation/7-1-diagnosing-poor-outputs

#### 7.2 — Adjusting Your Approach Based on Feedback and Results
Adjust like an investigator: change one thing at a time, hold the test input fixed, judge against a criterion stated in advance, and treat a third failed round as a signal to re-diagnose.
URL: https://claudecertificationguide.com/ccao-f/learn/7-troubleshooting-optimisation/7-2-adjusting-from-feedback

#### 7.3 — Optimising Workflows for Efficiency and Effectiveness
Optimise from a measurement, not a feeling: move repeated overhead to where it exists once, batch related requests, name files so retrieval can find them, and decline the optimisation that costs more than it saves.
URL: https://claudecertificationguide.com/ccao-f/learn/7-troubleshooting-optimisation/7-3-optimising-workflows

---

## CCDV-F: Claude Certified Developer (Foundations)

Free study guide for the Claude Certified Developer (Foundations) exam, covering all 8 domains: Agents & Workflows, Apps & Integration, Claude Code, Eval & Debugging, Model Selection, Prompt & Context, Security & Safety, and Tools & MCPs. 25 lessons and 369 practice questions.

### Domain 1: Agents and Workflows (14.7%)

Agent and workflow architecture, agent construction with the Claude Agent SDK, and common agentic patterns and frameworks.

#### 1.1 — Agent Architecture
When a workflow beats an agent and when it doesn't, how supervisor hierarchies delegate to subagents, and why context isolation is both the constraint and the payoff.
URL: https://claudecertificationguide.com/ccdv-f/learn/1-agents-workflows/1-1-agent-architecture

#### 1.2 — Agent Construction with Claude
The bare tool-use loop, the Claude Agent SDK's max_turns and max_budget_usd limits with their error_max_turns and error_max_budget_usd results, deployment models, and hooks as deterministic guardrails.
URL: https://claudecertificationguide.com/ccdv-f/learn/1-agents-workflows/1-2-agent-construction-claude

#### 1.3 — Agent Patterns and Frameworks
The five canonical workflow patterns and how to tell them apart, memory and context management as agent design decisions, and what LangGraph, Pydantic AI and Strands each lead with.
URL: https://claudecertificationguide.com/ccdv-f/learn/1-agents-workflows/1-3-agent-patterns-frameworks

### Domain 2: Applications and Integration (33.1%)

A third of the exam on its own: Claude API mechanics, software engineering foundations, application design, and configuration management.

#### 2.1 — Understanding Requirements
The business-to-functional-to-infrastructure derivation chain, why a latency target is never a functional requirement, and how rate limits and context windows become requirement inputs with documented numbers behind them.
URL: https://claudecertificationguide.com/ccdv-f/learn/2-applications-integration/2-1-understanding-requirements

#### 2.2 — Systems Life Cycle
Which life cycle phase owns which activity for a Claude application, why frameworks change cadence rather than concerns, and the deprecation clock that makes model migration a scheduled operations task.
URL: https://claudecertificationguide.com/ccdv-f/learn/2-applications-integration/2-2-systems-life-cycle

#### 2.3 — Claude API Mechanics
The Messages API wire shapes the exam tests exactly: request and response structure, stop_reason values, tool blocks, streaming events, vision, thinking, caching mechanics, batch versus realtime, vendor surfaces and the error table.
URL: https://claudecertificationguide.com/ccdv-f/learn/2-applications-integration/2-3-claude-api-mechanics

#### 2.4 — Software Engineering Foundations
HTTP and JSON as the substrate, bounded concurrency against documented rate limits, prompts as source code, the review bar for AI-assisted changes, and refactoring with a safety net.
URL: https://claudecertificationguide.com/ccdv-f/learn/2-applications-integration/2-4-software-engineering-foundations

#### 2.5 — Claude Application Design
The API as a blank slate versus the product layers of claude.ai and Claude Code, instruction placement, XML content boundaries, schema-first output contracts, session hygiene and plugin management as design decisions.
URL: https://claudecertificationguide.com/ccdv-f/learn/2-applications-integration/2-5-claude-application-design

#### 2.6 — Configuration Management
Every behaviour-bearing artefact versioned: the tracked-versus-local file boundary, what a model identifier actually resolves to, prompt versioning with environment parity, plugin pinning and secrets hygiene.
URL: https://claudecertificationguide.com/ccdv-f/learn/2-applications-integration/2-6-configuration-management

### Domain 3: Claude Code (3.1%)

Claude Code's core components and features, the CLAUDE.md hierarchy, repository initialisation, and settings.json configuration.

#### 3.1 — Claude Code Operation
The five components as files on disk, why CLAUDE.md files concatenate while settings.json obeys a precedence chain, session re-entry and the /clear versus /compact call, headless runs with claude -p, and the permission-mode axis.
URL: https://claudecertificationguide.com/ccdv-f/learn/3-claude-code/3-1-claude-code-operation

### Domain 4: Eval, Testing, and Debugging (2.6%)

Debugging and error handling for Claude applications: error identification, recovery strategies, trace analysis, and problem isolation.

#### 4.1 — Debugging and Error Handling
Isolating the integration layer from model output, the documented status-to-error-type map and the deterministic/transient split, stop_reason as a diagnostic on successful responses, errors that arrive after a 200, and matching the recovery to the class you established.
URL: https://claudecertificationguide.com/ccdv-f/learn/4-eval-testing-debugging/4-1-debugging-error-handling

### Domain 5: Model Selection and Optimisation (16.8%)

LLM fundamentals, supporting technical concepts, model capability trade-offs, and cost and token management.

#### 5.1 — LLM Fundamentals
Next-token generation and what follows from it, tokens measured rather than estimated, the context window as one shared budget with two distinct overflow behaviours, sampling and non-determinism, and which model option fixes which problem.
URL: https://claudecertificationguide.com/ccdv-f/learn/5-model-selection-optimisation/5-1-llm-fundamentals

#### 5.2 — Technical Fundamentals
The substrate a Claude application sits on: the raw HTTP and JSON wire, what the official SDK wraps and what it emphatically does not change, and which transport belongs at which layer when request/response, server-sent events and websockets are all on the table.
URL: https://claudecertificationguide.com/ccdv-f/learn/5-model-selection-optimisation/5-2-technical-fundamentals

#### 5.3 — Model Selection and Trade-offs
The durable cost and capability ordering behind the model families, selection as constraint-matching rather than tier dogma, the effort lever that usually beats a tier change, and the lifecycle states and breaking behaviour changes that make a migration a code change.
URL: https://claudecertificationguide.com/ccdv-f/learn/5-model-selection-optimisation/5-3-model-selection-trade-offs

#### 5.4 — Cost and Token Management
Pricing the token counts in usage before optimising anything, per-feature attribution that makes a saving provable, the asymmetric economics of prompt caching, where a checkpoint belongs on a growing conversation, and why max_tokens is a cap rather than a cost dial.
URL: https://claudecertificationguide.com/ccdv-f/learn/5-model-selection-optimisation/5-4-cost-token-management

### Domain 6: Prompt and Context Engineering (11%)

Context and memory management, prompt engineering principles and methods, and structured output handling.

#### 6.1 — Context Engineering
Bloat versus drift and why a bigger window fixes neither, tool-output pruning with clear_tool_uses_20250919, compaction and the preservation contract, and context isolation through subagents.
URL: https://claudecertificationguide.com/ccdv-f/learn/6-prompt-context-engineering/6-1-context-engineering

#### 6.2 — Prompt Engineering
System versus user placement and mid-conversation system messages, instruction clarity and few-shot examples, output constraints after prefill and sampling controls were removed, and input sanitisation as ordinary hygiene.
URL: https://claudecertificationguide.com/ccdv-f/learn/6-prompt-context-engineering/6-2-prompt-engineering

#### 6.3 — Output Handling
Structured outputs via output_config.format and strict tool use, why a guaranteed schema still needs a validator, checking stop_reason before the payload, and scepticism towards confident output.
URL: https://claudecertificationguide.com/ccdv-f/learn/6-prompt-context-engineering/6-3-output-handling

### Domain 7: Security and Safety (8.1%)

AI application security, guardrails and safe deployment, hooks as safety controls, and identity, secrets, and key management.

#### 7.1 — AI Application Security
The two prompt-injection threat models, delivering untrusted content as JSON-encoded tool_result blocks, screening with output_config, least privilege over the tool surface, prompt leak, PII retention, and the confused deputy.
URL: https://claudecertificationguide.com/ccdv-f/learn/7-security-safety/7-1-ai-application-security

#### 7.2 — Guardrails and Safe Deployment
Guardrail layering by enforcement point, the Usage Policy as an engineering obligation, least privilege and identity as secure-by-design, failing closed when a check is unavailable, and staged rollout with red-teaming.
URL: https://claudecertificationguide.com/ccdv-f/learn/7-security-safety/7-2-guardrails-safe-deployment

#### 7.3 — Claude Hooks
Hooks as deterministic control on the safety path: PreToolUse blocking with exit 2, the permissionDecision JSON form, where a hook sits in the permission evaluation order, and the limits worth stating out loud.
URL: https://claudecertificationguide.com/ccdv-f/learn/7-security-safety/7-3-claude-hooks

#### 7.4 — Identity, Secrets, and Key Management
Why a shipped client is public, the backend boundary that fixes it, key expiration and Workload Identity Federation, the exposure playbook in order, and attribution as the thing that makes an anomaly investigable.
URL: https://claudecertificationguide.com/ccdv-f/learn/7-security-safety/7-4-identity-secrets-key-management

### Domain 8: Tools and MCPs (10.6%)

Tool implementation, MCP server development, and choosing among built-in tools, custom tools, Skills, and MCPs.

#### 8.1 — Tool Implementation
What a tool description must say, why disambiguation is written rather than inferred, the informative tool_result with is_error that prevents fabrication, and where client and server tools actually execute.
URL: https://claudecertificationguide.com/ccdv-f/learn/8-tools-mcps/8-1-tool-implementation

#### 8.2 — MCP Server Development
Why one MCP server replaces four hand-written integrations, the host-client-server roles, the three server primitives and who controls each, and how deployment topology picks the transport.
URL: https://claudecertificationguide.com/ccdv-f/learn/8-tools-mcps/8-2-mcp-server-development

#### 8.3 — Agentic Customisation
Choosing between built-in tools, custom tools, Skills and MCP servers: what each one is, why a Skill cannot substitute for access, and the decision sequence that resolves a stem listing several requirements at once.
URL: https://claudecertificationguide.com/ccdv-f/learn/8-tools-mcps/8-3-agentic-customisation

---

## Blog

### Exam Guide or Current Docs: Which Answer Does the Claude Exam Want?
The Claude exam guides date from July 2026 and the product has moved on. What Anthropic says about the guide, why exams lag, and which answer to pick.
URL: https://claudecertificationguide.com/blog/exam-guide-vs-current-docs
Published: 2026-10-15

### Claude Code Certification: Does Anthropic Offer One? (2026)
Anthropic has no standalone Claude Code certification. Here's which of its exams test Claude Code, which one tests it most, and what the free courses give you.
URL: https://claudecertificationguide.com/blog/claude-code-certification
Published: 2026-09-28

### Claude Certified Architect (Foundations): Exam Guide and Blueprint
What the Claude Certified Architect (Foundations) exam actually tests: five domains, the scenario-based format, who it's for, and how to prepare for free.
URL: https://claudecertificationguide.com/blog/claude-certified-architect-foundations-exam-guide
Published: 2026-07-24

### Claude Certified Architect (Professional): Exam Guide and Blueprint
The Claude Certified Architect (Professional) exam: seven domains, senior audience, governance and lifecycle focus, and how it differs from Foundations.
URL: https://claudecertificationguide.com/blog/claude-certified-architect-professional-exam-guide
Published: 2026-07-24

### Claude Certified Associate (Foundations): Exam Guide and Blueprint
The Claude Certified Associate (Foundations) exam explained: the no-code credential for business users, its seven domains, who it's for, and how it's weighted.
URL: https://claudecertificationguide.com/blog/claude-certified-associate-foundations-exam-guide
Published: 2026-07-24

### Claude Certified Developer (Foundations): Exam Guide and Blueprint
The Claude Certified Developer (Foundations) exam: eight domains, 25 weighted sub-skills, who it's for, and why Applications and Integration dominates.
URL: https://claudecertificationguide.com/blog/claude-certified-developer-foundations-exam-guide
Published: 2026-07-24

### Anthropic's Claude Certifications: All Four Exams Explained (2026)
Anthropic's Claude Certification Program now spans four exams. Here's what each one tests, who it's for, and how to choose the right track.
URL: https://claudecertificationguide.com/blog/new-claude-certifications-2026
Published: 2026-07-24

### Building AI Agents with Claude: Orchestration Patterns That Actually Work
A practical guide to building AI agents with Claude. Covers the agentic loop lifecycle, single-agent vs multi-agent patterns, the coordinator/specialist architecture, the Claude Agent SDK, and what Domain 1 of the certification exam actually tests.
URL: https://claudecertificationguide.com/blog/building-ai-agents-claude-orchestration-patterns
Published: 2026-03-31

### Claude vs ChatGPT for Developers: An Honest Technical Comparison
A balanced, technical comparison of Claude and ChatGPT for developers. Covers API design, context windows, tool use, structured output, coding performance, pricing, and where each model genuinely excels.
URL: https://claudecertificationguide.com/blog/claude-vs-chatgpt-developers-comparison
Published: 2026-03-31

### How to Configure Claude Code with CLAUDE.md: The Complete Guide
A practical guide to configuring Claude Code projects with CLAUDE.md files. Covers the three-level hierarchy, modular organisation with @ path imports, path-specific rules, hooks, MCP servers, and what the certification exam actually tests.
URL: https://claudecertificationguide.com/blog/configure-claude-code-claude-md-guide
Published: 2026-03-31

### Claude Code Plugins: What They Are and How to Use Them
A practical guide to Claude Code Plugins. Covers the /plugin command, plugin.json structure, bundling skills, hooks, subagents and MCP servers, installing plugins, and the plugin marketplace.
URL: https://claudecertificationguide.com/blog/claude-code-plugins-guide
Published: 2026-03-24

### Claude Sonnet 4.6: What's New for Architects
A breakdown of Claude Sonnet 4.6 for solution architects. Covers adaptive reasoning, the 1M context window beta, improved agentic search, reduced token consumption, and the Claude Haiku 3 deprecation.
URL: https://claudecertificationguide.com/blog/claude-sonnet-4-6-whats-new
Published: 2026-03-24

### Structured Outputs Are Now GA on the Claude API
Structured outputs have moved from beta to GA on the Claude API. Covers what changed, how JSON schema validation works, practical examples with tool_use and direct schema enforcement, and what it means for the exam.
URL: https://claudecertificationguide.com/blog/claude-structured-outputs-ga
Published: 2026-03-24

### Context Window Management in Claude: Avoiding the Summarisation Trap
How to manage Claude's context window without losing critical data. Covers the progressive summarisation trap, tool result trimming, the lost-in-the-middle effect, and persistent case facts.
URL: https://claudecertificationguide.com/blog/context-window-management-claude-summarisation-trap
Published: 2026-03-18

### How to Pass the Claude Certified Architect Exam: Complete Study Guide
A practical, no-filler guide to the Claude Certified Architect (Foundations) exam. Covers all 5 domains, weighted study strategy, common traps, and the resources that actually matter.
URL: https://claudecertificationguide.com/blog/how-to-pass-claude-certified-architect-exam
Published: 2026-03-18

### MCP Servers Explained: How the Model Context Protocol Works
A practical guide to the Model Context Protocol (MCP). Explains how MCP servers, clients, and transports work — with runnable code examples and architectural patterns.
URL: https://claudecertificationguide.com/blog/mcp-servers-explained-model-context-protocol
Published: 2026-03-18

### Prompt Engineering for Claude: System Prompts, Few-Shot, and Output Validation
A hands-on guide to prompt engineering with Claude. Covers system prompt design, few-shot examples, structured output, and output validation — with runnable code and exam-tested patterns.
URL: https://claudecertificationguide.com/blog/prompt-engineering-claude-system-prompts-few-shot
Published: 2026-03-18