The Claude Certified Architect – Foundations (CCA-F) exam isn't weighted evenly. Its five domains each carry a different share of your score, and knowing those weights is the most efficient way to plan your study time. Spread your hours evenly across all five and you'll over-invest in the lightest domain while under-preparing for the heaviest. This guide breaks down the exact weightings and turns them into a concrete study schedule.
The five domains and their weights
| Domain | Weight | Approx. questions (of 60) |
|---|---|---|
| Agentic Architecture & Orchestration | 27% | ~16 |
| Claude Code Configuration & Workflows | 20% | ~12 |
| Prompt Engineering & Structured Output | 20% | ~12 |
| Tool Design & MCP Integration | 18% | ~11 |
| Context Management & Reliability | 15% | ~9 |
The exam has 60 scenario-based questions over 120 minutes, with a passing score of 720 out of 1000. You're presented with 4 of 6 possible scenarios, and questions use both multiple choice and multiple response formats.
Allocate your study time by weight
The simplest, most effective rule: allocate study hours in proportion to each domain's weight. For a 40-hour study plan:
| Domain | Weight | Hours (of 40) |
|---|---|---|
| Agentic Architecture & Orchestration | 27% | ~11 |
| Claude Code Configuration & Workflows | 20% | ~8 |
| Prompt Engineering & Structured Output | 20% | ~8 |
| Tool Design & MCP Integration | 18% | ~7 |
| Context Management & Reliability | 15% | ~6 |
What each domain covers
Agentic Architecture & Orchestration (27%) — your biggest priority
The heaviest domain. It covers designing multi-agent systems, task decomposition, hub-and-spoke orchestration models, agentic loops, subagent coordination, and reliability patterns for autonomous systems. Because it's the largest share of the exam, give it the most study time. Focus on when to use a single agent versus multiple, how to structure orchestration, and how to handle a subagent failure without silently returning incomplete output.
Claude Code Configuration & Workflows (20%)
Covers CLAUDE.md hierarchies, custom slash commands, and integrating Claude Code into CI/CD pipelines. Hands-on practice pays off here — actually configuring a CLAUDE.md file and building a slash command teaches this faster than reading about it.
Prompt Engineering & Structured Output (20%)
Covers enforcing reliability through JSON schemas, few-shot techniques, and validation retry loops. This domain is about producing predictable, machine-parseable output under real constraints, not casual prompting.
Tool Design & MCP Integration (18%)
Covers designing Model Context Protocol (MCP) servers, managing tool boundaries, and writing precise tool descriptions to prevent reasoning overload. Internalize the principle of minimal tool footprint: give an agent exactly the tools a task requires, and no more.
Context Management & Reliability (15%)
The lightest domain, but don't skip it. It covers preserving long context, managing handoff patterns, confidence calibration, and prompt caching and compaction for cost and reliability. Even at 15%, that's roughly 9 questions — more than enough to decide a pass.
Adjust for your own weak spots
Domain weight tells you the exam's emphasis; it doesn't tell you where you are weakest. The best approach is to start from the weight-proportional split above, then shift hours toward any domain where you're less confident. The fastest way to find those gaps is a diagnostic that scores you by domain — then you know exactly where your study hours will do the most good.
Bottom line
Study in proportion to the weights: Agentic Architecture & Orchestration (27%) first, then Claude Code Configuration and Prompt Engineering (20% each), then Tool Design & MCP Integration (18%), then Context Management & Reliability (15%) — adjusting toward your weak domains as a diagnostic reveals them. That's how you turn limited study time into the highest expected score.
Ready to put this into practice? Take the free 10-question readiness diagnostic to find your weakest domain in a few minutes, then build a personalised day-by-day schedule with the CCA Study Plan Generator. When you're ready to test your progress, the full 60-question timed simulation gives you a domain-by-domain score breakdown to guide what's next.