Three practice-style CCA-F mock exams covering the Claude Agent SDK, Claude Code, the Claude API, and MCP — with guided review, domain-by-domain scoring, and a study roadmap built from a real May 2026 pass journey.
CCA-F prep is still early. Learners piece together docs, video walkthroughs, scattered blog posts, and questionable question banks. The serious candidates want a focused, ethical practice environment with structured review.
This product is a clean study system: three practice-style mock exams, guided review, domain-based scoring, and a study roadmap. Built by someone who actually sat the exam.
Five parts. Three timed mock exams under realistic conditions. Guided review that explains the reasoning behind each answer. A study roadmap that walks you through the published domains. And weak-area recommendations after every attempt.
The published CCA-F blueprint splits scored content into five weighted domains across the Claude Agent SDK, Claude Code, the Claude API, and Model Context Protocol (MCP). Every sit draws four of six realistic production scenarios — customer support agents, code generation, multi-agent research, developer productivity, CI/CD integration, and structured data extraction. All three mocks mirror the same shape.
A focused, full-bleed exam interface — timer, question palette, flagging, and submit-with-confirmation. Built to remove the noise so you can practice the thinking.
Paid material is only available inside your account. Mock exams aren't provided as downloadable PDFs and print/export controls are intentionally omitted. The paid prep portal is a focused workspace, not a static library.
No subscription. No tiers. No upsells. A single one-time pass to the full ccaf.ai Field Manual — all three mock exams, guided review, domain scoring, study roadmap, and any updates we ship during the launch period.
This product does not contain real CCA-F exam questions, leaked content, copyrighted exam items, or unauthorized exam material. It is independent preparation material designed to help you practice concepts, timing, and reasoning.
The Claude Certified Architect — Foundations (CCA-F) was introduced by Anthropic in March 2026 as the company's first proctored, architecture-level credential — and the first formal benchmark for what production-grade work looks like across agent design, orchestration, and operational reliability. I earned the certification in May.
CCA-F is not a prompting assessment; it is a systems-design assessment. The exam evaluates the decisions that most often determine whether an agentic system succeeds or fails in production: scoping tool surfaces to minimize hallucinated calls, identifying when the Model Context Protocol is the appropriate abstraction versus when a lightweight custom integration is the better choice, composing subagents without context contamination, selecting the right model tier for each workload (knowing when Haiku is sufficient and when Opus justifies its cost), architecting Claude Code workflows that remain stable across long-horizon tasks, and applying the evaluation, caching, and guardrail practices that separate prototypes from dependable production systems.
The three mock exams provided here are designed to reflect that orientation. Each is scenario-based and reasoning-first, with no reused or leaked exam content. The objective is not to memorize answers, but to develop the judgment the certification is built to measure.