CCA-F Free Study Resources: The Official Anthropic Materials + CertLand Practice Pack

The Anthropic Claude Certified Architect — Foundations (CCA-F) is the first certification from an AI-native company that tests whether you can actually build production applications — not just prompt a chatbot. It covers five technical domains: agentic architecture, tool design & MCP integration, Claude Code configuration, prompt engineering, and context management.

Passing requires hands-on knowledge of the Claude Agent SDK, the Messages API, the Model Context Protocol, and real deployment workflows. The good news: Anthropic provides strong official documentation for all of it, and it is all free.

This post maps the official free resources to each exam domain, tells you exactly which documentation sections are most exam-relevant, and shows you how to combine them with CertLand's 606-question practice bank for the fastest path to passing.

Anthropic SkillJar: Free Official Courses

Anthropic's learning platform (anthropic.skilljar.com) hosts the official preparation courses for the CCA-F. All courses are free. You need to create an account to track your progress and access the course certificates, but none of the content is paywalled.

Here is what is available and which exam domains each course covers:

How to use SkillJar effectively: Don't treat these as passive video courses. The exam tests whether you can make architectural decisions, not whether you watched the videos. Work through each module, then immediately do 20–30 practice questions on the same topic in CertLand before moving on. This active recall loop is far more effective than watching alone.

Time estimate: The full SkillJar learning path takes approximately 6–8 hours to complete. Budget more time if you are pausing to take notes.

Official Documentation by Exam Domain

The CCA-F is rooted in Anthropic's own documentation. Anthropic publishes everything at docs.anthropic.com — API reference, concept guides, cookbook examples, and architecture patterns. This is the authoritative source for everything on the exam.

Below is a map of the most exam-relevant documentation sections for each domain. Not everything in the docs is exam-relevant; this cuts through the noise.

Domain 1: Agentic AI Architecture (27%)

The heaviest domain. Focus on understanding the orchestrator/subagent model, how tool loops work, and how to design agents that recover from errors gracefully.

What to focus on: The exam tests architectural judgment — given a scenario, which agent pattern is correct? Study the difference between sequential vs parallel subagent execution, when to use a subagent pool vs a single orchestrator, and what happens when a tool call fails mid-loop (retry vs abort vs fallback).

Common exam trap: Questions often have two plausible multi-agent designs. The correct answer is usually the one that minimizes unnecessary context propagation and handles errors at the right abstraction layer.

Domain 2: MCP Integration (18%)

The Model Context Protocol is Anthropic's standard for connecting LLMs to external tools and data sources. The exam tests whether you understand the protocol well enough to design MCP servers correctly.

What to focus on: Know the difference between Tools (actions), Resources (data), and Prompts (templates) — this distinction appears on the exam constantly. Understand when to use each, how resources are referenced by URI, and how MCP handles authentication at the server boundary.

Common exam trap: Questions about "which MCP primitive should you use for X" — if the use case is reading existing data (a file, a database row, an API result), that is a Resource. If it is performing an action with side effects, that is a Tool.

Domain 3: Claude Code (20%)

Claude Code is Anthropic's agentic CLI. The exam tests your ability to configure it correctly for production use — not just basic CLI usage.

What to focus on: Know CLAUDE.md scope rules cold — which CLAUDE.md takes precedence when there are multiple (project > user > enterprise)? Know which hook fires when (PreToolUse fires before the tool executes; PostToolUse fires after). Know what --allowedTools does vs --disallowedTools.

Read our complete Claude Code CLI reference guide for a comprehensive map of every command, flag, and configuration option.

Domain 4: Prompt Engineering (20%)

This domain covers the craft of writing effective prompts for production use — not just basic prompting, but structured output, tool use patterns, few-shot examples, and chain-of-thought.

What to focus on: The exam tests structured output and tool use heavily. Know when to use tool_choice: "any" vs "auto". Know how to enforce JSON output. Know how few-shot examples interact with system prompts. Read our Domain 4 deep dive for the full exam-relevant treatment.

Domain 5: Context Management (15%)

The smallest domain by weight, but one that trips up many candidates because the questions require understanding of implementation details rather than just concepts.

What to focus on: Prompt caching rules appear consistently on the exam. Know that the cache TTL is 5 minutes for ephemeral caches. Know which content types are cacheable (system prompts, long documents passed as user turns) and what invalidates the cache. Know the tradeoffs between in-context memory and external memory for long-horizon tasks.

CertLand: Practice Questions + AI Coach

Reading documentation builds conceptual knowledge. Practice questions reveal what you actually understand vs what you think you understand. These are different things, and the CCA-F exam will surface the gap.

CertLand's CCA-F preparation system gives you:

Putting It All Together: A 2-Week Resource Plan

Here is how to sequence the free resources above for a focused 14-day preparation:

If you use the AI Coach on CertLand, the reading schedule above is built in automatically — the coach surfaces the relevant study post before each domain block, so you never have to figure out what to read next.