How many questions are on the Architect Foundations exam?

The CCAR-F exam has 60 items in 120 minutes. Items are multiple-choice and multiple-response, and each item states how many responses to select.

What is the passing score for CCAR-F?

Scaled score of 720 on a scale of 100–1,000.

How much does the Architect exam cost and how long is it valid?

The exam fee is $125 USD. The credential is valid for 12 months from the date the credential is awarded.

Which domains does the Architect exam cover?

5 domains: 1. Agentic Architecture & Orchestration; 2. Tool Design & MCP Integration; 3. Claude Code Configuration & Workflows; 4. Prompt Engineering & Structured Output; 5. Context Management & Reliability.

CCAR-F

Claude Certified Architect – Foundations

Prove you can design agentic systems: orchestration, tools and MCP, structured output, and reliability under real context limits.

Exam facts

Credential: Claude Certified Architect – Foundations
Exam code: CCAR-F
Number of items: 60
Item format: Multiple-choice and multiple-response items; each item states how many responses to select
Exam structure: 4 scenarios drawn from a bank of 6
Time limit: 120 minutes
Delivery: Proctored: online proctored and/or test center, per program policy
Passing score: Scaled score of 720 on a scale of 100–1,000
Exam fee: $125 USD
Validity period: 12 months from the date the credential is awarded

Official domain blueprint

Listed in domain order, with the published weight and task statements for each.

Domain 1: Agentic Architecture & Orchestration

27% of the exam — Loops, delegation, subagents, and knowing when not to use an agent.

Design and implement agentic loops for autonomous task execution
Orchestrate multi-agent systems with coordinator-subagent patterns
Configure subagent invocation, context passing, and spawning
Implement multi-step workflows with enforcement and handoff patterns
Apply Agent SDK hooks for tool call interception and data normalization
Design task decomposition strategies for complex workflows
Manage session state, resumption, and forking

Domain 2: Tool Design & MCP Integration

18% of the exam — Tool surfaces, descriptions, error shapes, and MCP server boundaries.

Design effective tool interfaces with clear descriptions and boundaries
Implement structured error responses for MCP tools
Distribute tools appropriately across agents and configure tool choice
Integrate MCP servers into Claude Code and agent workflows
Select and apply built-in tools (Read, Write, Edit, Bash, Grep, Glob) effectively

Domain 3: Claude Code Configuration & Workflows

20% of the exam — Project configuration, permissions, hooks, and repeatable dev workflows.

Configure CLAUDE.md files with appropriate hierarchy, scoping, and modular organization
Create and configure custom slash commands and skills
Apply path-specific rules for conditional convention loading
Determine when to use plan mode vs direct execution
Apply iterative refinement techniques for progressive improvement
Integrate Claude Code into CI/CD pipelines

Domain 4: Prompt Engineering & Structured Output

20% of the exam — Schemas, validation, and output contracts other systems can consume.

Design prompts with explicit criteria to improve precision and reduce false positives
Apply few-shot prompting to improve output consistency and quality
Enforce structured output using tool use and JSON schemas
Implement validation, retry, and feedback loops for extraction quality
Design efficient batch processing strategies
Design multi-instance and multi-pass review architectures

Domain 5: Context Management & Reliability

15% of the exam — Compaction, retrieval, retries, and graceful degradation.

Manage conversation context to preserve critical information across long interactions
Design effective escalation and ambiguity resolution patterns
Implement error propagation strategies across multi-agent systems
Manage context effectively in large codebase exploration
Design human review workflows and confidence calibration
Preserve information provenance and handle uncertainty in multi-source synthesis

Exam scenarios

Four of these six scenarios are drawn for any given sitting; items are grouped under the scenario they belong to.

Customer Support Resolution Agent

An Agent SDK support agent handling returns, billing disputes and account issues through custom MCP tools, targeting 80%+ first-contact resolution while knowing when to escalate.

Code Generation with Claude Code

Claude Code across generation, refactoring, debugging and docs, wired into a team workflow with custom slash commands, CLAUDE.md config and plan mode vs direct execution.

Multi-Agent Research System

A coordinator delegating to specialised subagents for web search, document analysis, synthesis and report generation, producing comprehensive cited reports.

Developer Productivity with Claude

Agent SDK tooling that explores unfamiliar codebases, explains legacy systems and automates repetitive work using built-in tools plus MCP servers.

Claude Code for Continuous Integration

Claude Code in a CI/CD pipeline running automated reviews, generating tests and giving pull-request feedback that is actionable with few false positives.

Structured Data Extraction

Extraction from unstructured documents with JSON-schema validation, high accuracy, graceful edge-case handling and downstream system integration.

How to prepare for CCAR-F

1. Self-assess against every objective
Work down the blueprint below and mark each published task statement as confident, shaky, or unknown. The exam writes its items directly against these objectives, so an honest self-assessment is the whole study plan in miniature.
2. Learn the item format, not just the content
Items are multiple-choice and multiple-response, and each one states how many responses to select. Every question is a realistic situation asking for the soundest judgement — practise choosing between four defensible-looking options, not reciting definitions.
3. Work the six scenarios, not isolated facts
The exam presents 4 scenarios drawn from a bank of 6. Read each scenario in the guide and reason through the architecture you would ship for it — the items are all framed inside these production contexts.
4. Build with the Agent SDK, MCP, and Claude Code
Hands-on time across all three: coordinator and subagent delegation, lifecycle hooks, MCP tool and resource interfaces, CLAUDE.md files, custom slash commands, and plan mode versus direct execution.
5. Default to deterministic enforcement
A recurring judgement on this exam: when a business rule must hold, programmatic enforcement beats prompt-based instruction. Practise recognising when prompting is not a reliability guarantee.
6. Drill structured output and context strategy
Design JSON schemas with enum and extensible fields, write extraction prompts, and plan for long documents, multi-turn sessions, and multi-agent handoffs — including error handling and human-in-the-loop escalation.

Frequently asked questions

How many questions are on the Architect Foundations exam?: The CCAR-F exam has 60 items in 120 minutes. Items are multiple-choice and multiple-response, and each item states how many responses to select.
What is the passing score for CCAR-F?: Scaled score of 720 on a scale of 100–1,000.
How much does the Architect exam cost and how long is it valid?: The exam fee is $125 USD. The credential is valid for 12 months from the date the credential is awarded.
Which domains does the Architect exam cover?: 5 domains: 1. Agentic Architecture & Orchestration; 2. Tool Design & MCP Integration; 3. Claude Code Configuration & Workflows; 4. Prompt Engineering & Structured Output; 5. Context Management & Reliability.

Try 40 free Architect questions with full explanations, then move on to the timed sets.

Free questions See pricing