TUTORIALS
Claude Certified Architect โ Foundations (CCA-F): a complete self-study tutorial
A seven-part self-study series for Anthropic's first official Claude certification. Covers the five exam domains, six scenarios, pass tips from real candidates, and a 30-question practice exam.
A seven-part series for the Claude Certified Architect โ Foundations exam. Roughly 60 minutes of reading, plus a 30-question practice exam. Designed to be followed in order.
Anthropic launched the Claude Certified Architect โ Foundations (CCA-F) certification on 12 March 2026. It is a proctored, scenario-based exam: 60 multiple-choice questions in 120 minutes, $99 USD, pass at 720 / 1000. Most candidates with hands-on Claude experience need 15โ20 hours to prepare; complete newcomers should plan 30โ40 hours.
This series covers every domain on the published syllabus, with concrete facts, design patterns, and the kind of trade-offs the exam actually tests.
The seven parts
Part 1: Exam Overview and How to Study
What the Claude Certified Architect Foundations exam is, who it's for, the five domains and their weightings, and a three-week study plan. (about 3 min)
Part 2: Agentic Architecture and Orchestration (27% of the exam)
The biggest domain on the exam: single agents, supervisor-worker, plan-and-execute, evaluator-optimiser, subagents, escalation patterns, and common anti-patterns. (about 3 min)
Part 3: Claude Code Configuration and Workflows (20%)
CLAUDE.md hierarchy, .claude/rules/ with path scoping, skills with context: fork, plan vs direct mode, and the CI/CD patterns the exam tests. (about 3 min)
Part 4: Prompt Engineering and Structured Output (20%)
Specificity over wishes, few-shot examples, tool_use schemas, validation-retry loops, anti-hallucination patterns, and when to use the Message Batches API. (about 3 min)
Part 5: Tool Design and MCP Integration (18%)
Tool description hygiene, structured error responses with categories, MCP roles and primitives, transports (stdio vs Streamable HTTP), and tool scoping per agent. (about 3 min)
Part 6: Context Management and Reliability (15%)
Context windows, compaction, scratchpad files, subagents for token-heavy reads, bounded retries, idempotency keys, and the production escalation rule. (about 3 min)
Part 7: Exam Scenarios, Tips from People Who Passed, and All Resources
The six exam scenarios, eight tips from candidates who scored above 720, and a comprehensive list of free and paid prep resources. (about 11 min)
How to use this series
- Week 1 โ Read it through. Work top to bottom. Take the practice exam below the article (or via
/exams) as a diagnostic. - Week 2 โ Build the patterns. Pick two of the six exam scenarios and build working prototypes in Claude Code or against the Claude API.
- Week 3 โ Mock-exam loop. Re-take the practice exam under timed conditions (60 minutes for 30 questions โ exam pace). Book the real exam when you score above 80% twice in a row.
Take the practice exam
A 30-question practice exam matching the format and weighting of the real CCA-F is available in the exams hub. Each question has a single correct answer; explanations cover why the right answer is right and why each wrong answer is wrong.
This is an independent study guide. It is not affiliated with, endorsed by, or sponsored by Anthropic.
Written by
Mohamed AL-Kaisi
Editor-in-chief of the Data & AI Hub.