Claude Certified Developer – Foundations Concept Library

Explore concepts across all 8 exam domains. Each domain maps to the Anthropic Claude Certified Developer – Foundations (CCDV-F) certification and contains task statements with atomic concepts, glossary terms, and practice questions.

38 task statements85 concepts

Applications & Integration

Translate business requirements into a Claude application's design, operate that design through the systems life cycle and the Messages API's concrete mechanics, build it with sound software-engineering practices, match instruction design to whichever Claude interface is in play, and keep every layer of configuration versioned and reproducible.

8 task statements·17 concepts·102 questions

33.1% of exam

Model Selection

Understand how LLMs actually work (tokens, context, sampling, non-determinism), how SDKs relate to the underlying REST API, and how to choose among Claude's model tiers and cost/latency levers -- caching and batching -- to run applications economically.

6 task statements·14 concepts·54 questions

16.8% of exam

Agents & Workflows

Decide how much autonomy a system needs -- choosing between predefined workflow patterns and open-ended agents -- and build that decision with the Claude Agent SDK or a custom harness, manager/subagent hierarchies, hooks, and abstraction frameworks.

4 task statements·9 concepts·48 questions

14.7% of exam

Prompt & Context

Write clear, well-placed instructions that reliably steer individual calls, curate everything in the context window so Claude has the right information and nothing that distracts it, and consume Claude's output safely through structured formats, validation, and defensive parsing.

5 task statements·11 concepts·36 questions

11% of exam

Tools & MCPs

Extend Claude beyond text by defining function-calling tools, building MCP servers that expose reusable tools/resources/prompts, and choosing correctly among built-in tools, custom tools, Skills, and MCP servers for a given requirement.

3 task statements·8 concepts·36 questions

10.6% of exam

Security & Safety

Build Claude applications that are secure by design: isolate untrusted content and apply least privilege to defeat prompt injection and jailbreaks, prevent PII and data leakage, layer guardrails with deterministic hooks, and manage identity, secrets, and API keys correctly.

4 task statements·8 concepts·24 questions

8.1% of exam

Claude Code

Operate Claude Code as an agentic coding tool -- its core component types, the CLAUDE.md memory hierarchy, settings.json as the permissions/behavior control surface, and the session, headless, streaming, and auto-mode operating features.

4 task statements·9 concepts·12 questions

3.1% of exam

Eval & Debugging

Identify the type of a Claude application failure, isolate whether it originates in the integration layer or the model output, use trace analysis to find the earliest faulty step in a multi-step workflow, and select and verify the right recovery strategy.

4 task statements·9 concepts·6 questions

2.6% of exam

Exam Domain Weights

Applications & Integration

33.1%

Model Selection

16.8%

Agents & Workflows

14.7%

Prompt & Context

11%

Tools & MCPs

10.6%

Security & Safety

8.1%

Claude Code

3.1%

Eval & Debugging

2.6%