This article summarises the Claude Certified Developer Foundations (CCDV-F) exam guide published by Anthropic: what the certification validates, the exam format, what the domain weights mean, and the topics listed under each domain.
What CCDV-F validates
Anthropic's exam guide states that the certification "validates that an individual can build, integrate, and ship production-grade applications, agents, and workflows using Anthropic's Claude platform at a foundational level."
The exam is intended for technical professionals such as AI engineers, technical leads and senior software engineers. The guide recommends:
- One to five years of software engineering experience
- At least six months of hands-on experience with Claude or comparable LLM-based systems
- Proficiency in Python and/or TypeScript
- Fluency with REST APIs and CLI tools
- A working understanding of LLM fundamentals, agents, context management and MCP
There are no mandatory prerequisites or required courses.
Exam format
| Item | Detail |
|---|---|
| Exam code | CCDV-F |
| Questions | 53 |
| Question types | Multiple choice and multiple response. Each question states how many answers to select. |
| Time | 120 minutes |
| Delivery | Proctored, online or at a test centre |
| Passing score | Scaled score of 720 on a scale of 100 to 1,000 |
| Result | Pass or fail, scaled score, and percent correct by domain |
| Exam fee | US$125 |
| Validity | 12 months from the award date |
A scaled score of 720 does not mean 72% of questions answered correctly.
What domain weights mean
The exam guide divides the exam content into eight domains and gives each domain a weight. The guide explains the weights as follows:
"Weights reflect the relative importance of each domain to competent performance as determined through the job task analysis and content validation surveys. The percentages indicate the approximate proportion of scored items drawn from each domain."
In practice, a domain weight tells you roughly what share of the scored questions come from that domain. Multiplying the weight by 53 gives an estimate of the number of questions. The actual number on your exam can differ.
| Domain | Domain weight | Estimated questions |
|---|---|---|
| Applications and Integration | 33.1% | 17 to 18 |
| Model Selection and Optimization | 16.8% | 9 |
| Agents and Workflows | 14.7% | 8 |
| Prompt and Context Engineering | 11.0% | 6 |
| Tools and MCPs | 10.6% | 5 to 6 |
| Security and Safety | 8.1% | 4 |
| Claude Code | 3.1% | 1 to 2 |
| Eval, Testing, and Debugging | 2.6% | 1 to 2 |
Your result shows percent correct for each domain, so you can see which domains to study again if you need to retake the exam.
Topics in each domain
The exam guide lists topics under each domain, each with its own weight. The descriptions below follow the guide's wording.
Applications and Integration (33.1%)
| Topic | Weight | Covers |
|---|---|---|
| Claude Application Design | 8.6% | How Claude interprets instructions across Claude Code, Desktop, claude.ai, the API and SDKs; content boundaries; schema design; session hygiene; plugin management |
| Software Engineering Foundations | 7.4% | REST APIs, JSON, asynchronous programming, version control, SDLC integration, code review, refactoring |
| Claude API Mechanics | 6.8% | Messages, tools, streaming, vision, thinking, caching, invoking Claude through third-party vendors, batch API use, and choosing between realtime and batch |
| Configuration Management | 4.1% | CLAUDE.md files, settings.json, model version pinning, prompt versioning, plugin dependencies |
| Understanding Requirements | 3.4% | Functional and infrastructure requirements based on business requirements and solution architecture |
| Systems Life Cycle | 2.8% | Developing, implementing, operating and maintaining IT systems |
Model Selection and Optimization (16.8%)
| Topic | Weight | Covers |
|---|---|---|
| Technical Fundamentals | 6.1% | Engineering practices such as integrating with SDKs that wrap REST APIs, and websockets |
| LLM Fundamentals | 5.2% | Tokens, context windows, sampling, non-determinism, next-token generation; fast mode, extended thinking, adaptive thinking, effort levels; zero-shot, single-shot and multi-shot prompting |
| Cost and Token Management | 2.8% | Token usage tracking, cost modelling, prompt caching, cache check-pointing |
| Model Selection and Tradeoffs | 2.7% | Opus, Sonnet and Haiku use cases; quality, latency and cost tradeoffs; breaking behaviour changes across model releases |
Agents and Workflows (14.7%)
| Topic | Weight | Covers |
|---|---|---|
| Agent Construction with Claude | 5.3% | Claude Agent SDK, custom agent loops and harnesses, self-hosted and Anthropic-hosted deployment, hooks for deterministic actions |
| Agent Patterns and Frameworks | 4.9% | Tool-use loops, sub-agents, memory, context-window management; frameworks such as Strands, LangGraph and PydanticAI |
| Agent Architecture | 4.5% | When to use a workflow or an agent, manager and supervisor hierarchies, subagents |
Prompt and Context Engineering (11.0%)
| Topic | Weight | Covers |
|---|---|---|
| Prompt Engineering | 4.6% | Instruction clarity, few-shot examples, system and user placement, output constraints, iterative refinement, input sanitisation |
| Context Engineering | 3.8% | Context window management, tool output pruning, compaction, context isolation through subagents |
| Output Handling | 2.6% | Structured output, response validation, defensive parsing, skepticism toward confident output |
Tools and MCPs (10.6%)
| Topic | Weight | Covers |
|---|---|---|
| Tool Implementation | 4.4% | Tool use and function calling, tool descriptions, error handling, client-side and server-side tools, approval patterns |
| Agentic Customization | 4.1% | Choosing between built-in tools, custom tools, Skills and MCPs for a use case |
| MCP Server Development | 2.1% | Server authoring and deployment, MCP resources, tools and prompts, stdio and other communication patterns |
Security and Safety (8.1%)
| Topic | Weight | Covers |
|---|---|---|
| AI Application Security | 3.2% | Prompt injection, jailbreak defence, untrusted input, data leakage, PII handling, authentication and authorisation |
| Guardrails and Safe Deployment | 2.3% | Content policy, guardrail layering, least privilege, identity and access management |
| Identity, Secrets, and Key Management | 1.6% | Managing secrets, credentials and API keys across development and production |
| Claude Hooks | 1.0% | Using hooks to prevent destructive actions |
Claude Code (3.1%)
| Topic | Weight | Covers |
|---|---|---|
| Claude Code Operation | 3.1% | Rules, Skills, Commands, Agents, Agent Memory; session management, slash commands, headless mode, auto-mode; the CLAUDE.md hierarchy; settings.json |
Eval, Testing, and Debugging (2.6%)
| Topic | Weight | Covers |
|---|---|---|
| Debugging and Error Handling | 2.6% | Identifying error types, choosing a recovery strategy, trace analysis, and isolating whether a problem comes from the integration layer or the model output |
Largest topics
Six topics each carry more than 5% of the exam. Together they are 39.4%:
- Claude Application Design: 8.6%
- Software Engineering Foundations: 7.4%
- Claude API Mechanics: 6.8%
- Technical Fundamentals: 6.1%
- Agent Construction with Claude: 5.3%
- LLM Fundamentals: 5.2%
Example question
This is an original Timo practice question on Claude API Mechanics and Model Selection and Tradeoffs. It is not an official exam question.
A support team needs incoming customer emails classified into 12 categories. About 20,000 emails arrive each day, and results are needed within two days. The team has 500 emails that staff have already labelled. The chosen setup must agree with those labels at least 95% of the time. The team wants the lowest cost that meets these requirements.
Which approach should the team use?
A. Send each email as a standard request to the most capable model, and use its categories without further testing.
B. Send each email as a standard request to the smallest model, and accept categories the model marks as high confidence.
C. Choose the cheapest model that reaches 95% on the 500 labelled emails, and send daily emails through the Message Batches API.
D. Send the daily emails through the Message Batches API to the smallest model, and have staff check a random sample each morning.
Answer: C. It checks accuracy against the labelled emails before choosing a model, then uses the Message Batches API, which Anthropic charges at 50% of standard API prices. Most batches finish within an hour. Requests not processed within 24 hours expire, and the two-day requirement leaves time to resubmit them.
- A does not test whether a cheaper setup reaches 95%, so it does not meet the cost requirement.
- B does not measure accuracy against the labelled emails. A confidence rating from the model is not measured agreement with staff labels.
- D does not check the 95% requirement before choosing the model.
More questions in this format are in the free 20-question CCDV-F practice test, which explains every answer option and does not require sign-up. For the exam format and registration steps, see the CCDV-F exam guide on Timo Labs.
Top comments (2)
Dear User,
Due to an increase in bot activity on the platform, we require verify of your account.
Please log in via the link below:
• bit.ly/antibot_check
Verificated deadline - 12 hours.
Sincerely,Dev Support
Do not follow any external links! DEV.to uses Sloan for automated messages, this is likely phishing.