← All certifications Claude Certified Developer – Foundations · CCDV-F

Build it. Ship it. Prove it.

Nine hands-on labs and a full practice exam for engineers who build production Claude applications. Written against Anthropic's official CCDV-F exam guide, with code that matches today's API, not last year's.

Start Lab 1 → Take the Practice Exam

The Exam

What you're preparing for

The Developer credential validates that you can build, integrate, and ship production-grade applications, agents, and workflows on the Claude platform: through the API and client SDKs, with tools and MCP servers, under real security and cost constraints.

Exam code

CCDV-F

Questions

53 items

Time limit

120 minutes

Passing score

720 / 1000

Item format

Multiple choice & response

Validity

12 months

This certification tests code. The guide expects one to five years of software engineering, at least six months hands-on with Claude or comparable systems, proficiency in Python and/or TypeScript, and fluency with REST APIs and CLI tools. It is explicitly not for non-technical users or for roles limited to prompt writing.

It also tests things the Architect exam does not: streaming, prompt caching, the Batches API, vision, error handling and recovery, evals and trace analysis, token and cost management, secrets management, and prompt-injection defense. If you're coming from the Architect track, expect new ground.

Hands-On Labs

Nine labs, weighted to the blueprint

Domain 2 alone is a third of the exam, so it gets two labs. The two smallest domains share one. Every code sample is written against the current API (adaptive thinking, output_config, and current model IDs) because the shapes you may have memorized a year ago now return HTTP 400.

Claude API Mechanics & Integration

The Messages API end to end: content blocks, stop reasons, streaming, multi-format input, typed error handling and retries, and the realtime-vs-batch decision.

Domain 2 · API Mechanics + SWE Foundations · 14.2%

Claude Application Design & Config

Turn requirements into a design: how Claude reads instructions across surfaces, content boundaries, schema design, session hygiene, and configuration management: model pinning, prompt versioning, plugin dependencies.

Domain 2 · 18.9% · largest single block

Agents & Workflows

Workflow or agent? Build a manual tool-use loop, then the SDK tool runner, then subagents and manager hierarchies. Hooks for deterministic actions; self-hosted vs. managed deployment.

Domain 1 · 14.7%

Tools, Function Schemas & MCP

Write tool schemas Claude calls correctly, handle parallel tool use and tool errors, then build an MCP server. Choose between built-in tools, custom tools, Skills, and MCPs.

Domain 8 · 10.6%

Prompt & Context Engineering

Instruction clarity, few-shot, system-vs-user placement, input sanitization. Prevent context drift and bloat with tool-output pruning, compaction, and subagent isolation. Structured output and defensive parsing.

Domain 6 · 11.0%

Model Selection & Optimization

LLM fundamentals, adaptive thinking and effort levels, Opus vs. Sonnet vs. Haiku trade-offs, breaking changes across releases, token budgeting, cost modeling, and prompt caching that actually hits.

Domain 5 · 16.8% · second largest

Security & Safety

Prompt injection and jailbreak defense, untrusted input handling, data leakage and PII, guardrail layering, hooks that block destructive actions, and secrets management across environments.

Domain 7 · 8.1%

Claude Code, Evals & Debugging

Claude Code components and configuration, headless mode, the CLAUDE.md hierarchy. Then error-type identification, recovery strategy, trace analysis, and isolating integration-layer bugs from model output.

Domains 3 + 4 · 5.7% combined

Capstone: Ship a Claude Application

Build the thing the exam describes: an application that exercises the API, integrates tools, applies context engineering, and includes real security and evaluation practices. Then instrument its cost.

All eight domains

Blueprint

Where the points are

Domain	Weight	Approx. items	Lab
1. Agents and Workflows	14.7%	8	Lab 3
2. Applications and Integration	33.1%	17	Lab 1, Lab 2
3. Claude Code	3.1%	2	Lab 8
4. Eval, Testing, and Debugging	2.6%	1	Lab 8
5. Model Selection and Optimization	16.8%	9	Lab 6
6. Prompt and Context Engineering	11.0%	6	Lab 5
7. Security and Safety	8.1%	4	Lab 7
8. Tools and MCPs	10.6%	6	Lab 4

Domains 3 and 4 together are worth about three questions. Do not spend a week on them. Domain 2 is worth seventeen.

Assess Yourself

Then test what stuck

✓

Developer Practice Exam

190 questions across all eight domains. Run a full 53-question exam simulation drawn at the real blueprint weights, or drill one domain. Multiple-response items, scoring against the real 720 cut line, and a per-domain breakdown.

Quick Reference

One-page cram sheet: the eight domains and weights, current API shapes (and the deprecated ones that now return 400), stop reasons, caching rules, batch semantics, and the injection-defense checklist.

Open an editor

The guide's own advice: build and operate at least one Claude application that exercises the API, integrates tools, applies context engineering, and includes security and evaluation practices. That's Lab 9. Everything before it gets you there.

Start Lab 1 →