2 packs: 15% off with code BUY2. 3 or more: 20% off with code BUY3.

Anthropic CCDV-F Exam Practice Questions, practice question pack
ANTHROPIC · CCDV-F

Anthropic CCDV-F Exam Practice Questions

99% pass rateUpdated 1 Nov 202696 questions2 PDFs, instant download

US$39

Every question with the correct answer and why the other options are wrong.

Download the free sample

Card, Apple Pay, Google Pay or PayPal. Download links are on your order page and in your email the moment you pay.

99% pass rate. Pass or your money back.Fail the exam after using this pack and we refund it, or swap it for any other pack. How the guarantee works
Category:
TRY BEFORE YOU BUY

Three of the 96 questions in this pack

Question 1

You are starting a new Claude application and have a small set of well-labeled examples that demonstrate the desired output format. You want to use these examples to guide Claude's behavior.

How would you guide the application's behavior?

  1. Embed the examples in a database and retrieve them at runtime as reference material for the team.
  2. Use zero-shot prompting and rely on the model's general capability to produce the desired output format consistently across all incoming requests.
  3. Train a custom model on the labeled examples before deployment so that the application does not need to include the examples in any prompt during operation.
  4. Use multi-shot prompting by including the labeled examples in the prompt so Claude can match the desired output format on each request.
Show answer and explanation

Correct answer: D. Use multi-shot prompting by including the labeled examples in the prompt so Claude can match the desired output format on each request.

Multishot prompting places a few well-labeled examples directly in the prompt so Claude can infer the desired output format and imitate it on every request. Anthropic's prompt engineering guidance recommends examples as one of the most reliable ways to steer format, structure, and tone, and notes that they are more effective than abstract instructions alone. Because the examples travel with each request, no training run, extra infrastructure, or model change is needed, and the behavior stays easy to adjust by editing the examples.

Why the other options are wrong

  • A. A database of examples kept as team reference material never reaches the model, so it cannot influence the format of Claude's output.
  • B. Zero-shot prompting gives Claude no demonstrations, so it discards the labeled examples and makes consistent adherence to a specific format less reliable.
  • C. Training a custom model is unnecessary for a small set of examples and removes the flexibility of simply showing Claude the examples in the prompt.

Question 2

Your Claude application's content policy specifies categories of content it should not produce under any circumstance. The application currently has no mechanism to enforce this policy, and content matching these categories is appearing in the application's output.

How would you enforce the content policy?

  1. Enhance the system prompt to contain explicit instructions for the categories to avoid, complete with examples of each category. Treat the strengthened prompt as the primary enforcement mechanism for the application's content policy across all responses.
  2. Remove the content policy entirely and let any output reach users during normal operation, accepting whatever content the application produces in response to incoming traffic.
  3. Add deterministic output filtering that checks responses against the content policy before they reach users.
  4. Move enforcement to users by asking them to report content policy violations after the violating content has already reached them in the application's responses.
Show answer and explanation

Correct answer: C. Add deterministic output filtering that checks responses against the content policy before they reach users.

A policy that must hold under any circumstance needs a check that does not depend on the model's judgment, so deterministic output filtering applied before responses reach users is the right control. Prompt instructions guide Claude but cannot guarantee compliance, while a filter or classifier on the output enforces the rule every time and can block or replace violating content. Anthropic's guardrail guidance likewise recommends screening and post-processing of outputs as a layer on top of prompting, and this layer also gives the team a record of what was blocked.

Why the other options are wrong

  • A. A strengthened system prompt improves the odds of compliance but is probabilistic, so on its own it cannot guarantee that prohibited content never appears.
  • B. Removing the policy abandons the requirement entirely and lets the prohibited content continue to reach users.
  • D. Relying on user reports only finds violations after the harm has occurred and shifts enforcement onto the people the policy is meant to protect.

Question 3

A new Claude model release includes performance improvements for several reasoning tasks but has changed the format of its responses to system prompts that use multi-section instructions. Your application uses multi-section system prompts heavily. Initial evaluation on the application's actual workload shows the new model performs 8 percent better on reasoning tasks but produces malformed output on roughly 3 percent of requests because of the format change. The team is debating whether to upgrade.

How would you decide?

  1. Upgrade and add a downstream validation step that catches the 3 percent malformed output before it reaches users, treating the validation step as the team's mitigation for the format change.
  2. Adapt the application's system prompt to the new model's format expectations and re-evaluate, then upgrade only if the adapted prompt eliminates the malformed output while preserving the reasoning improvements.
  3. Stay on the previous model permanently to avoid the malformed output rate and any future format changes that subsequent model releases might introduce in the application.
  4. Upgrade immediately, because the 8 percent reasoning improvement outweighs the 3 percent malformed output rate across the application's typical request distribution.
Show answer and explanation

Correct answer: B. Adapt the application's system prompt to the new model's format expectations and re-evaluate, then upgrade only if the adapted prompt eliminates the malformed output while preserving the reasoning improvements.

A format change in how a new model reads multi-section system prompts is usually something prompt adaptation can fix, so the sound approach is to adjust the prompt to the new model, then re-run the evaluation on the application's real workload. The upgrade should proceed only if the malformed output is eliminated and the reasoning gains remain. Anthropic's migration guidance recommends testing prompts against the new model and tuning them, because prompts written for one model often need small changes to get the best behavior from another.

Why the other options are wrong

  • A. Catching bad output downstream treats the symptom while leaving roughly 3 percent of requests failing and adding retry cost and latency, when the cause may be fixable in the prompt.
  • C. Staying on the previous model permanently gives up the reasoning gains and eventually runs into the model's deprecation and retirement schedule.
  • D. Upgrading immediately ignores that malformed output breaks the application's responses, and a reasoning gain does not offset requests that fail outright.

See all 24 free questions Get the full pack, US$39

How much of the Anthropic CCDV-F bank is free?

96 questions in the bank

Each step includes the one before, so buying only adds new questions.

  • You are here10questionsFree sampleThree on this page, all ten as a PDF.Download the sample
  • 2425% of the bankPractice pageFree on the site, every answer explained.Go to the practice page
  • 48the first half of the packHalf PackFree by email, as one PDF, every answer explained.Get the Half Pack
  • 96100%, the whole bankFull pack, US$39Plus the questions-only PDF, monthly updates, refund if you fail.Get the full pack
ABOUT THIS PACK

What’s in the Anthropic CCDV-F pack

  • 96 real questions on building with Claude, from the API and SDK to agents
  • Answers and explanations for every question, including the wrong options
  • A questions-only PDF for timed practice runs
  • Instant delivery by email the moment you check out
  • Free monthly updates for as long as the exam is live
  • Pass or your money back

The exam itself costs US$125 per attempt before partner discounts. This pack is US$39, paid once, and refunded if you fail.

What makes the CCDV-F hard

The CCDV-F exam is open only to employees of Claude Partner Network organisations. You register with a partner email address, you must be 18 or older, and you need a government ID on the day. If your organisation is not a partner, confirm your eligibility with Anthropic before buying this or any other CCDV-F preparation.

This is a developer exam, so the questions are about what you would actually build. A prompt keeps returning the wrong format and you choose between few-shot examples, a schema, or a prefilled response. A tool is called with the wrong arguments. A hook should block an action before it runs. A bill is climbing and you pick the right model for the job. Most options are plausible; one fits the constraint in the scenario.

This pack has 96 real questions covering the Claude API and SDK, prompt and context engineering, tool use and MCP, Skills and hooks, Claude Code configuration, agents and workflows, structured output, model selection and cost, security and safety, and evaluation and debugging, so the question style is familiar before exam day.

Studying for the architect exam as well? See the Anthropic CCAR-F pack, and the related Anthropic CCAO-F pack and Anthropic CCAR-P pack.

About the exam

Claude Certified Developer, Foundations (CCDV-F) is one of Anthropic's Claude certifications for the Claude Partner Network. It is delivered by Pearson VUE, online with OnVUE or at a test centre, in English and Japanese. The exam has multiple-choice questions and scenario-based multiple-response questions.

Anthropic does not publish the number of questions or the domain weights in its certification FAQ, so this page does not state them.

120 minutes to answer, pass mark 720 on a scaled 100 to 1,000, US$125 per attempt before partner discounts.

Questions before you buy

How quickly do I receive it?

Straight away. The full PDF and a questions-only copy are emailed to you the moment your payment goes through, and the same links are on your order page.

Are updates included?

Yes. The pack is updated every month for as long as the exam is live, and updates are free for everyone who has bought it.

What if I fail the exam?

We refund the pack, or swap it for any other pack if you would rather keep studying. Sit the exam 7 to 30 days after buying, then send your official score report within 7 days of the exam date, as set out in the refund policy.

What is your pass rate?

99% of customers who prepare with a CertStash pack pass. If you do not, we refund the pack, or swap it for another pack.

Can I share it with colleagues?

Each purchase is licensed to one person. For a team, school or training organisation, email support@certstash.com for a licence that fits.

Related products