Skip to content
← All certifications

Anthropic · CCDV-F · Foundations

Claude Certified Developer – Foundations

Building with the Claude API: applications and integration, agents and workflows, tools and MCP, model selection, prompt and context engineering, security and evals. Currently for Claude Partner Network staff.

Official exam page ↗

What the exam covers

  • Agents and Workflows15%
  • Applications and Integration33%
  • Claude Code3%
  • Eval, Testing, and Debugging3%
  • Model Selection and Optimization17%
  • Prompt and Context Engineering11%
  • Security and Safety8%
  • Tools and MCPs10%

Weights are the vendor's published exam guide.

Step 1 · Assess

Start with a 20-question diagnostic

Free, no account. Drawn across every domain in the exam’s own weights, so the score means what the real thing will. Treat it like the exam and don’t look anything up — the map is only useful if it’s true.

Already preparing? Sign in

About the Claude Certified Developer – Foundations (CCDV-F) exam

The Claude Certified Developer – Foundations is Anthropic's certification for engineers who build applications on Claude. It tests whether you can design and ship a working integration with the Claude API: calling the model correctly, handling tools and agents, choosing the right model, engineering prompts and context, and keeping the result secure.

The exam guide, effective July 2026, has eight domains with uneven weights. Applications and integration is a third of the paper on its own, followed by model selection and optimisation, agents and workflows, prompt and context engineering, and tools and MCP. Security and safety, Claude Code, and evaluation and debugging make up the rest.

At the time of writing the certification is open only to staff of organisations in the Claude Partner Network. Anthropic publishes free prep courses; this practice is ours, written against the public exam guide.

The exam at a glance

Questions
53 questions, multiple choice and multiple response
Time
120 minutes
Pass mark
720 on a scale of 100 to 1,000
Exam fee
About US$125
Valid for
12 months; renew before it lapses with Anthropic's free renewal assessment
Where you sit it
Pearson VUE, online proctored or at a test centre
Prerequisites
None, but you must work for a Claude Partner Network organisation to register

As published in the vendor’s exam guide at the time of writing. Fees vary by country and change; the vendor’s page is the source of truth.

What is on the exam, domain by domain

  1. Agents and Workflows

    When to build a workflow and when an agent, and how to structure multi-step work.

    15% of the paper

  2. Applications and Integration

    The largest domain, about a third. Designing a Claude application, the Messages API and its mechanics, configuration, requirements and the software life cycle around an integration.

    33% of the paper

  3. Claude Code

    Using Claude Code in a development workflow.

    3% of the paper

  4. Eval, Testing, and Debugging

    Testing Claude-based features and finding why they misbehave.

    3% of the paper

  5. Model Selection and Optimization

    Choosing a model for the job and tuning for quality, latency and cost.

    17% of the paper

  6. Prompt and Context Engineering

    Writing system prompts and instructions, and managing what goes into the context window.

    11% of the paper

  7. Security and Safety

    Prompt injection, data handling, guardrails and safe defaults for an application.

    8% of the paper

  8. Tools and MCPs

    Defining tools, handling tool calls and results, and connecting systems through the Model Context Protocol.

    10% of the paper

How to prepare

  1. Build something end to end before the exam: a small app that calls the Messages API, uses a tool, and handles errors and retries. The integration questions assume you have done it.
  2. Read Anthropic's developer documentation on messages, tool use, prompt caching and streaming, and learn what each is for.
  3. Spend a third of your time on applications and integration; it is a third of the paper.
  4. Learn the Model Context Protocol at the level of servers, tools and resources, and when you would use it rather than defining tools directly.
  5. Take Anthropic's free prep course, then practise papers here until you clear your target twice running.

Sample questions

Three questions in the exam’s style, written by Square 1 against the published domains. They are ours, not the vendor's. The answer and the reasoning are under each one; the practice papers on this page draw from a bank written the same way, with an explanation on every question.

  1. Question 1 · Applications and Integration

    Your application receives an overloaded or rate-limit error from the Claude API during a traffic spike. What is the most robust way to handle it?

    1. Retry with exponential backoff, and respect any retry-after information the API returns
    2. Retry immediately in a tight loop until the request finally succeeds
    3. Fail the user's request at once and ask them to try again later
    4. Switch every future request to a smaller model for the rest of the day

    Answer: A. Transient capacity and rate-limit errors are meant to be retried after a delay, backing off exponentially and honouring any retry-after signal. Retrying in a tight loop makes the congestion worse and is likely to keep failing.

  2. Question 2 · Tools and MCPs

    Claude keeps choosing the wrong one of two similar tools in your application. What is the first thing to improve?

    1. The tool names and descriptions, so each says clearly what it does and when to use it
    2. The temperature setting, lowering it so tool choice becomes more predictable
    3. The maximum number of output tokens, raising it so Claude can reason for longer
    4. The order of the tools in the request, putting the preferred tool first in the list

    Answer: A. Claude chooses tools from their names and descriptions, so vague or overlapping descriptions are the usual cause of wrong choices. Clear, distinct descriptions that say when to use each tool fix the root cause; sampling settings do not.

  3. Question 3 · Security and Safety

    Your application asks Claude to summarise web pages that users submit. What is the main risk to design against?

    1. Instructions hidden in a page that try to redirect Claude, known as prompt injection
    2. Pages that are longer than usual and so take a little more time to summarise
    3. Pages written in a language other than the one the user reads
    4. Pages that include images which the summary cannot fully describe

    Answer: A. Untrusted content can carry instructions meant to hijack the model, which is prompt injection. Treat fetched content as data, not instructions, and limit what the model can do with it; length and language are quality concerns, not security ones.

Questions people ask before they book

Who can take the Claude Certified Developer exam?
At the time of writing, only staff of organisations in the Claude Partner Network, registering with a partner email address. Anthropic has not announced a date for opening it to everyone.
How technical is it?
It is a developer exam. You need to understand the Claude API, tool use, agents and prompt engineering well enough to design and debug an integration, though you will not write code during the exam.
Which domain matters most?
Applications and integration, at about a third of the paper. Model selection and agents come next. Claude Code and evaluation are small shares, so do not over-invest in them.
How long is the certification valid?
Twelve months, renewable before it lapses with a free renewal assessment; after a lapse you sit the full exam again.
Are these the real exam questions?
No. Every question here is ours, written against Anthropic's public exam guide. Anthropic's exam content is confidential, and practising on leaked questions would break its exam policy.
Claude Certified Developer – Foundations (CCDV-F) practice questions and exam guide | Square 1 AI