Claude Certification Blog
Ten CCDV-F worked scenarios, explained option by option
These ten CCDV-F worked scenarios are original items written to the official Exam Guide v1.0, spread across six of the eight Developer domains by weight, each with the key and a stated reason why every option holds or fails.
The ten CCDV-F worked scenarios below are not exam items and were not written from any exam item. Each is a short scenario with a measured symptom and a constraint, then four options, or five for a select-two. Nine are single response and one is select-two, the two formats the guide describes. Answer every item cold, record the misses in the map below, and open the reason for every option afterwards. Three items sit in Applications and Integration because that domain is a third of the paper.
The official CCDV-F exam guide is the authority on format and weights; the Claude Certified Developer - Foundations page carries the blueprint and six more open items. The reading method these items reward is in the Claude certification sample questions hub, and what official material exists is in the guide to Claude certification practice exams.
Why do these CCDV-F worked scenarios skip two domains?
Because the blueprint does. Applications and Integration is 33.1% of the exam, about 17 of 53 items; Eval, Testing, and Debugging is 2.6%, a single item, and Claude Code is 3.1%, two. Spread ten items by those weights and the two smallest domains round to zero, so the set covers the six domains that carry 94% of the paper; the free CCDV-F practice exam, all 53 items, carries the other two at their exam allocation. The omission states how few marks they are worth; it is not advice to skip them.
Applications and Integration: which layer actually changed?
Three items, because the domain is a third of the paper, on three published skills: API mechanics, where prompt caching pays only when the cached prefix is identical from the first token; software engineering foundations, where an idle CPU beside blocked threads is an I/O problem, not a capacity problem; and configuration management, where a floating model alias makes an upstream release an untested production change. The distractors share one pattern: a bigger knob where the fault is structural.
A 6,000-token system prompt has a cache_control breakpoint at its end, yet cache reads stay near zero. The prompt opens with the current timestamp and user ID, then static instructions. What fixes the cache?
Show the answer and the reason for every option
Answer: B
Caching matches an exact prefix from the first token. Objective O06, Claude API Mechanics.
A. A second breakpoint still sits after the changing first line, so no request ever shares a prefix with another.
B (correct). Correct. With the changing values after the breakpoint, every request shares the same static prefix and reads it from the cache.
C. Each prefix is unique because of the timestamp, so a longer lifetime keeps entries that nothing will ever match.
D. At 6,000 tokens the prompt is already long enough to cache. The changing first line is the problem.
A Python service waits about four seconds on each synchronous Claude API call. At 300 users all 32 threads block on network waits, with CPU under 10 percent. Which change fixes the cause?
Show the answer and the reason for every option
Answer: C
Network waits call for async code. Objective O07, Software Engineering Foundations.
A. More threads let more requests wait at once, but each still blocks a thread and the blocking design remains.
B. Only identical messages would hit such a cache, so most requests still block a thread for four seconds.
C (correct). Correct. Awaiting calls on an event loop lets one worker serve many requests while each waits on the network.
D. A faster model shortens each wait, but every call still blocks a thread for its whole duration.
With no deployment, a report service's output changed section order one Tuesday and parsing failed for six hours. The code names its model by a floating alias. What prevents a repeat?
Show the answer and the reason for every option
Answer: C
Pin the model and upgrade on evidence. Objective O09, Configuration Management.
A. Temperature affects sampling within one version. It cannot stop a new release from changing the output format.
B. Retrying does not fix a format change. The parser rejects the new order every time.
C (correct). Correct. With a dated version named in configuration, the model changes only on a deliberate, tested upgrade.
D. CLAUDE.md is configuration for Claude Code sessions. It does not govern a production service's prompt or model.
Model Selection and Optimization: is the extra spend buying anything?
Two items on the second-heaviest domain, both turning on a measurement the scenario already contains: an evaluation that shows no meaningful gap from extended thinking on a bounded classification, and a bill that more than doubled behind a service that logs nothing. The defensible option acts on the evidence in the stem, not on a general belief about reasoning or cheaper tiers. Watch the mechanism error too: the thinking budget must fit inside the output limit.
An email router choosing one of six queues runs extended thinking with a 10,000-token budget. Latency is 6.2 seconds against a 1.5-second target. Accuracy is 96.4 percent with thinking, 96.2 without. What should you change?
Show the answer and the reason for every option
Answer: D
Thinking is a per-call setting, spent where it pays. Objective O12, LLM Fundamentals.
A. A smaller model with thinking still spends time reasoning on a task the eval shows does not need it.
B. The evidence shows almost no accuracy gain, so asking the business to accept slower routing has no basis.
C. A lower cap with thinking on can cut the reply off before the answer, and the budget is still spent first.
D (correct). Correct. The eval shows thinking barely helps this call, and turning it off brings latency back within the target.
Your API bill rose 2.4 times in a month. The service serves 60 accounts and logs nothing per request. You must find which accounts and request types drive the rise. What comes first?
Show the answer and the reason for every option
Answer: A
Measure usage per account before changing anything. Objective O15, Cost and Token Management.
A (correct). Correct. Per-request token counts tagged by account and request type show exactly what drives the increase.
B. Splitting by traffic assumes every request costs the same, which hides the very differences you need to find.
C. A blanket cap changes behavior for every account before anyone knows which ones drive the increase.
D. Switching models changes cost and quality everywhere without finding out what caused the increase.
Agents and Workflows: what must happen on every run?
Two single-response items. The first is the compliance case the Agent SDK's hooks exist for: a step the model usually performs is not a step that always happens, and a compliance team that accepts no probabilistic control has ruled out every option that is still an instruction. The second is a context problem disguised as a runtime problem. Fanning the documents out to subagents, each with its own context and returning only a summary, fixes both symptoms at once. An external store is the closest miss: the findings survive, yet a single agent still reads every document, slowly.
A Claude Agent SDK coding agent must validate every file write against a schema. The prompt instruction to do so was skipped in 6 percent of runs, and compliance rejects probabilistic controls. What meets this?
Show the answer and the reason for every option
Answer: B
A required step belongs in a hook. Objective O02, Agent Construction with Claude.
A. Repeating the instruction may cut the miss rate, but the model still decides whether to follow it.
B (correct). Correct. The hook runs in code on every write and rejects failures, so the check never depends on the model.
C. A reviewing subagent is another model making a choice, so it is still a probabilistic control.
D. A stronger model may skip less often, but it still decides for itself whether to run the check.
A due-diligence agent reads about 300 documents in one session. Early findings drop out of the report once the window fills, and a run takes four hours. Which pattern fixes both problems?
Show the answer and the reason for every option
Answer: A
Isolate context per subagent and run in parallel. Objective O03, Agent Patterns and Frameworks.
A (correct). Correct. Each subagent reads its batch in a separate context and sends back a summary, and the batches can run in parallel.
B. Saved findings survive, but one session still reads all 300 documents, so the window fills and the run stays slow.
C. Frameworks run on the same model and window, so switching does not stop the window from filling.
D. Shorter reasoning saves a little space, but 300 documents still overflow one window and the run stays slow.
Security and Safety: what reduces the exposure in the logs?
One select-two item on a domain worth 8.1% of the exam. The exposure is not the model call but the copy of every prompt and completion that hundreds of staff can read for two years. The published skill is data leakage prevention and PII handling, and the two options that meet it act on the data and on the audience. The distractors act elsewhere: on the model's output, on the storage medium, on a credential nobody has exposed. Encryption at rest is a real control aimed at a threat the stem never describes.
A claims assistant's prompts include names, birth dates and medical notes. Each prompt and completion is logged in full for two years to a dashboard 300 staff can open. Which two changes reduce the exposure? (Select two.)
Show the answer and the reason for every option
Answer: B and D
Minimize stored PII and restrict who can read it. Objective O19, AI Application Security.
A. The prompts themselves contain the personal data and are logged in full, so a rule about replies changes nothing there.
B (correct). Correct. Redaction keeps raw names, dates and medical notes out of the logs, so the dashboard never holds them.
C. Encryption at rest protects the disks, but anyone with dashboard access still reads the data in full.
D (correct). Correct. Role-based access and an access record limit who can open the logs and show who did.
E. Key rotation protects the credential. The logs still hold raw personal data that 300 staff can open.
Tools, MCPs and output handling: where do data, identity and shape get decided?
Two single-response items from domains that together carry 21.6% of the paper. The MCP item is a transport decision: where the data lives, where credentials may live and whose identity each query must carry settle the choice between a local process and a remote server; the closest miss authenticates the server when the constraint was the person. The output item is the structured-output contract in miniature: a schema through a forced tool call, validation before insert and a retry path.
Staff will query a records system through an MCP server from Claude Desktop. It runs only in the data center, laptops hold no credentials, and queries must use each user's identity. Which deployment fits?
Show the answer and the reason for every option
Answer: D
Match the transport to where data and identity live. Objective O24, MCP Server Development.
A. A shared service account authenticates the server, not the person, so no query runs under the user's own identity.
B. A stdio server runs on the laptop, so the shared credential would sit on every machine, which is ruled out.
C. A nightly export puts a copy of the records on every laptop, and no query uses the user's own identity.
D (correct). Correct. Hosting the server next to the records, with per-user sign-in, keeps credentials off laptops and ties each query to a person.
Claude extracts incident reports as JSON for a nightly load. About 3 percent open with a sentence or have an invalid severity, crashing the job at the first bad record. Which approach loses no reports?
Show the answer and the reason for every option
Answer: D
Constrain the shape, validate it, retry failures. Objective O18, Output Handling.
A. Stripping text fixes a leading sentence, but an invalid severity still reaches the database and can crash the job.
B. An instruction makes good output more likely but enforces nothing, so some records will still be malformed.
C. Skipping a malformed record keeps the job running but loses that report, which the stem rules out.
D (correct). Correct. The schema and enum fix the shape and values, and validation with retry recovers the rest without losing a report.
How to score these CCDV-F questions: the miss map
Score exactly: a select-two item counts only when both letters match. Then write each miss into its domain row rather than adding up a total, because ten items cannot estimate a scaled score. What they can do is point. One miss in a domain means reread the skill named in that row in the official guide. Two or more means the domain link, which opens the free runner's untimed review filtered to that domain, before you sit the full timed paper.
| Domain | Items | Your misses | Objective to reread | Practise the domain |
|---|---|---|---|---|
| D1 Agents and Workflows | 2 | ___ of 2 | O02 Agent Construction with Claude; O03 Agent Patterns and Frameworks | Domain 1 in the free mock |
| D2 Applications and Integration | 3 | ___ of 3 | O06 Claude API Mechanics; O07 Software Engineering Foundations; O09 Configuration Management | Domain 2 in the free mock |
| D5 Model Selection and Optimization | 2 | ___ of 2 | O12 LLM Fundamentals; O15 Cost and Token Management | Domain 5 in the free mock |
| D6 Prompt and Context Engineering | 1 | ___ of 1 | O18 Output Handling | Domain 6 in the free mock |
| D7 Security and Safety | 1 | ___ of 1 | O19 AI Application Security | Domain 7 in the free mock |
| D8 Tools and MCPs | 1 | ___ of 1 | O24 MCP Server Development | Domain 8 in the free mock |
The wrong option is usually a bigger knob
Across these ten, the most tempting distractor makes an existing number larger: a longer cache lifetime, more threads, a higher tier, a stricter instruction. The key changes what breaks. Before you pick, ask whether the option alters the mechanism or only its size.
The set pairs with the two-week CCDV-F study plan, which schedules the domains in weight order, and the CCDV-F cheat sheet, whose rules these items exercise. The alternative, a memorised answer key, is covered in what Claude certification exam dumps actually cost you. The CCDV-F blueprint sets the weights used here. Checked against the official CCDV-F Exam Guide v1.0 (July 2026) on 17 September 2026.
Key takeaways
- Ten original items, six domains, by weight. Applications and Integration gets three because it is a third of the exam; the two smallest domains live in the full mock.
- Answer cold, then reveal. Reading the reasons before committing turns a diagnostic into a reading exercise.
- Score by domain, not by total. Ten items cannot estimate the scaled score; a miss pattern in one domain can name the skill to reread.
- Distrust the bigger knob. Longer, more, higher and stricter are the shape of most distractors here; the key changes the mechanism.
- Read the constraints as a checklist. Every distractor in the MCP and security items fails at least one stated constraint; the key fails none.
Now sit the full 53 under a clock
The free CCDV-F mock is 53 timed items across all eight domains with a reason for every option, no account needed. Signed in, the account adds 5 aligned timed forms, 265 items, plus drills by objective. Nothing reports a scaled score or a pass probability.
Start the free CCDV-F mockQuestions
Frequently asked
Short answers to the questions this post leaves open.
Are these real CCDV-F exam questions?
No. All ten were written fresh for this page from the published CCDV-F Exam Guide v1.0 and its domain skills, and none is drawn from, or paraphrased from, any exam form or any gated Cred Farmer form. Anthropic keeps its live items confidential; a page offering CCDV-F exam questions that claims otherwise is selling a breach.
How many questions are on the CCDV-F exam?
The official guide sets 53 items in 120 minutes across eight domains, mixing single-answer and multiple-response items, with every multiple-response stem stating how many options to select. The pass mark is a scaled 720 on a 100 to 1,000 scale; the fee is $125 per attempt before partner-tier discounts.
Is there a free CCDV-F mock exam?
Yes. The Cred Farmer runner at /practice/ccdv-f sits 53 original items in 120 minutes with no account, allocated across the eight domains by official weight, and explains every option after you submit. It reports a raw count by domain, never a scaled score, a pass verdict or a percentile.
Where can I find CCDV-F sample questions?
Three places, all free: the illustrative samples inside the official exam guide, the six open items on the Cred Farmer certification page, and the ten worked items here. All three follow the same eight-domain blueprint; only the guide is official. The retired official practice exam has not been replaced.
How should I score myself on these CCDV-F questions?
Count an item right only when your letters match the key exactly, so a select-two item with one correct pick is a miss. Write the misses into the map by domain and act on the row, not the total. Ten items say nothing about the scaled score, and no practice percentage converts to it.
Keep reading
Related posts
Not affiliated with, or endorsed by, Anthropic or Pearson VUE. Details are summarised from publicly published program information and can change. Always confirm against the official exam guide before booking.