Claude Certified Architect Foundations (CCA-F) Exam is Anthropic’s first official technical certification, launched March 12, 2026 as part of the Claude Partner Network. It validates the ability to design, deploy, and maintain production-grade applications using the Claude ecosystem, including the Claude Agent SDK, Claude Code, and the Model Context Protocol (MCP). This is an architecture-level credential — it tests system design decisions, not everyday Claude usage or prompt trivia.

WHO IT IS FOR:

• Solution architects building production applications with Claude

Recommended experience: at least 6 months of hands-on experience with the Claude API and Claude Code.

EXAM DOMAINS for Claude Certified Architect Foundations (CCA-F) Exam Questions 2026

Domain 1 – Agentic Architecture and Orchestration (27%) Designing multi-agent systems, task decomposition, hub-and-spoke orchestration models, agentic loops, subagent coordination, and reliability patterns for autonomous systems.

Domain 2 – Tool Design and MCP Integration (18%) Designing Model Context Protocol servers and clients, managing tool boundaries, SSE transport trade-offs, authentication patterns, and writing reliable tool schemas.

Domain 3 – Claude Code Configuration and Workflows (20%) CLAUDE markdown hierarchy, custom slash commands, CI/CD pipeline integration, persistent project context, and AI-first development workflows.

Domain 4 – Prompt Engineering and Structured Output (20%) JSON schema enforcement, few-shot techniques, validation retry loops, programmatic enforcement vs. prompt-based guidance.

Domain 5 – Context Management and Reliability (15%) CALM framework, prompt caching with cache_control breakpoints, conversation compaction, token estimation and budget management, multi-turn conversation design.

EXAM DETAILS

Format: 60 multiple-choice questions, scenario-based Duration: 120 minutes Passing Score: 720 (scaled score of 100-1,000) Exam Fee: USD 99 Languages: English Delivery: Online proctored or in-person testing center Structure: 6 production scenarios, 4 of which are randomly selected per sitting. All questions are anchored to those scenarios.

EXAM STYLE

This is not a documentation quiz. Every question places you inside a realistic production scenario — a broken agent loop, a poorly scoped tool schema, a system under load — and asks you to make the correct architectural call. Candidates who only read documentation passively tend to underperform. Scenario-based practice is essential.

PREPARATION

Official resources hosted on Anthropic Academy (Skilljar platform):

• Building with the Claude API (flagship course, 8+ hours)

Recommended prep time: 4-12 weeks depending on hands-on experience.