CCA Foundations study guide
Welcome to the community study guide for the Claude Certified Architect — Foundations (CCA) certification. This guide covers every domain tested on the exam, with task statements drawn directly from the official exam guide PDF.
The CCA does not test whether you know how to use Claude. It tests whether you can architect production-grade Claude applications — designing multi-agent systems, configuring Claude Code for teams, building reliable MCP integrations, and making the right engineering trade-offs under production constraints.
How to use this guide
- Start with the overview — understand the exam format, all 6 scenarios, and the 6-week study plan at /overview
- Read each domain — use the sidebar to navigate domain deep-dives in order of exam weight (Domain 1 → 2 → 3 → 4 → 5)
- Drill the anti-patterns — review the 7 canonical wrong answers before sitting any practice questions
- Practice — work through flashcards (concepts) then quiz (application)
- Target weaknesses — the quiz lets you filter by domain and difficulty level; drill your weak domains
Exam weight at a glance
| Domain | Topic | Weight |
|---|---|---|
| 1 | Agentic Architecture & Orchestration | 27% |
| 3 | Claude Code Configuration & Workflows | 20% |
| 4 | Prompt Engineering & Structured Output | 20% |
| 2 | Tool Design & MCP Integration | 18% |
| 5 | Context Management & Reliability | 15% |
Domains 1 + 3 alone make up 47% of the exam. If time is short, prioritize them.
The mental model that passes the exam
Most wrong answers on the CCA are architecturally plausible but miss a single key trade-off. The exam rewards one consistent mental model:
Deterministic > probabilistic. When the exam asks how to enforce ordering, compliance, or policy rules — the answer is always programmatic (hooks, prerequisites, schema validation), never prompt-based (instructions, few-shot examples). Prompts guide; code enforces.
Everything else flows from this:
stop_reasonover text parsing- Programmatic prerequisites over system prompt instructions
isRetryableflags over generic retry logic- JSON schema +
strict: trueover "ask Claude nicely" - Scoped tools over giving agents everything
Official preparation resources
| Resource | URL |
|---|---|
| Anthropic Academy (13 free courses) | anthropic.skilljar.com |
| Agent SDK docs | platform.claude.com/docs/en/agent-sdk/overview |
| Claude Code docs | code.claude.com/docs/en/overview |
| Tool use docs | platform.claude.com/docs/en/agents-and-tools/tool-use/overview |
| MCP documentation | modelcontextprotocol.io/docs |
| Official exam guide PDF | Download PDF |