Claude Certified Architect Foundations / CCA-F

A clearer path to Claude architecture.

The exam covers the decisions behind production Claude systems. See what it assesses, where to focus, and which official resources to use.

Explore the domains → Find study resources

Independent study overview. Algorims is not affiliated with or endorsed by Anthropic; trademarks belong to their owners.

60multiple-choice questions
120 minexam duration
5weighted domains
$125per attempt

01 / The overview

Know the exam before you study.

CCA-F is for practitioners at Anthropic Partner companies who already have hands-on experience with Claude. It assesses architectural judgment across agents, tools, prompts, workflows, and reliability.

The exam is delivered through Pearson VUE, online or at a test center. The published pass score is 720 on a scaled 100–1,000 range. You should be comfortable reading Python or TypeScript and working with JSON.

New to Claude? Start with Anthropic Academy ↗ before working through the domains below.

02 / Exam domains

Focus where it counts.

Published exam weights. Open a domain to see its task statements.

01 Agentic Architecture & OrchestrationAgentic loops, coordinator and subagent patterns, task decomposition, and session state. 27%
  1. Design and implement agentic loops
  2. Orchestrate multi-agent systems
  3. Configure subagent invocation and context passing
  4. Implement multi-step workflows with enforcement
  5. Apply Agent SDK hooks
  6. Design task decomposition strategies
  7. Manage session state, resumption, and forking
02 Tool Design & MCP IntegrationTool interfaces, structured errors, MCP servers, and tool access across agents. 18%
  1. Design effective tool interfaces
  2. Implement structured error responses for MCP tools
  3. Distribute tools across agents and configure tool choice
  4. Integrate MCP servers into Claude Code and agent workflows
  5. Select and apply built-in tools effectively
03 Claude Code Configuration & WorkflowsCLAUDE.md, reusable commands and skills, plan mode, and CI/CD workflows. 20%
  1. Configure CLAUDE.md files with hierarchy and modular organisation
  2. Create and configure custom slash commands and skills
  3. Apply path-specific rules for conditional convention loading
  4. Determine when to use plan mode vs direct execution
  5. Apply iterative refinement techniques
  6. Integrate Claude Code into CI/CD pipelines
04 Prompt Engineering & Structured OutputClear criteria, few-shot prompts, JSON schemas, validation, and review passes. 20%
  1. Design prompts with explicit criteria to reduce false positives
  2. Apply few-shot prompting for consistency and quality
  3. Enforce structured output using tool use and JSON schemas
  4. Implement validation, retry, and feedback loops
  5. Design efficient batch processing strategies
  6. Design multi-instance and multi-pass review architectures
05 Context Management & ReliabilityLong-running context, escalation, error propagation, and uncertainty. 15%
  1. Manage conversation context across long interactions
  2. Design escalation and ambiguity resolution patterns
  3. Implement error propagation strategies across multi-agent systems
  4. Manage context in large codebase exploration

03 / Practice

Study through real scenarios.

Use these as prompts to test how the domains work together. They are study examples, not official exam questions.

01

Support automation

Design an agent that resolves requests, calls tools, and escalates when confidence is low.

Domains 1 · 2 · 5
02

Developer workflow

Configure Claude Code, team conventions, and context management for codebase work.

Domains 3 · 5
03

Research orchestration

Coordinate specialist agents, tool access, source handling, and synthesis.

Domains 1 · 2 · 5
04

Productivity automation

Choose tool boundaries, prompts, and workflows for repeated knowledge work.

Domains 1 · 2 · 3
05

CI/CD assistance

Use structured output, review criteria, and non-interactive execution in pipelines.

Domains 3 · 4
06

Document extraction

Design schemas, validation, retries, and reliable context use for messy inputs.

Domains 4 · 5

04 / Resources

Start with the source.

Official registration, courses, and documentation, grouped by what you need next.

Register

Build foundations

Go deeper

The Anthropic Skilljar courses for Bedrock, Vertex AI, and the Claude API include foundation and advanced material. Check the official certification page for current access, pricing, and exam details.