Preparing for the Claude Certified Architect Foundations Exam

0.2.1 Lesson Overview

0.2.2 Install Node

0.2.3 Getting a Claude API Key

Student Exercises - Starter Projects and Solutions

0.2.4 Setting up VS Code

0.2.5 Installing Dependencies and Testing Setup

Welcome to Module 1

Module 1 Pre-Assessment

1.1.1 The Agentic Loop Overview

1.1.2 The Single Call Pattern

1.1.3 Introducing the Agentic Loop

1.1.4 Why this Matters Everywhere

1.1.5 The Loop Lifecycle

1.1.6 Sending the Request

1.1.7 The Response Shape

1.1.8 Stop Reason as Control Signal

1.1.9 The Stop Reason Table

1.1.10 Executing Tools and Tool Use Ids

1.1.11 Role Alternation and the User Role Trap

1.1.12 A Full 2 Iteration Example

1.1.13 What Happens Inside a Turn

Knowledge Check: The Agentic Loop Lifecycle

1.2.1 Implementing Your First Agentic Loop Overview

1.2.2 Finding the Starter Files

1.2.3 Review the Tool Definitions

1.2.4 Send your First Request

1.2.5 Execute the Tool Call

1.2.6 Close the Loop

1.2.7 Test Your Loop

1.3.1 Stop Reason Deep Dive - Lesson Overview

1.3.2 Max Tokens Truncation

1.3.3 Detecting and Recovering from Truncation

1.3.4 The isError Field

1.3.5 Retryable vs Terminal Errors

1.3.6 Error Handling in the Loop

1.3.7 Stop Sequence Defining Boundaries

1.3.8 Refusal and Pause Turn

1.3.9 Iteration Limits

1.3.10 The Robust Loop Template

Knowledge Check: stop_reason Tool Results, and Loop Termination

1.4.1 Anti Patterns in Loop Controle - Overview

1.4.2 Text Parsing Anti-Pattern

1.4.3 Silent Failure Anti-Pattern

1.4.4 Content Spiral Anti-Pattern

1.4.5 Patterns and Takeaways

Knowledge Check: Anti-patterns in Loop Control

1.5.1 Multi Agent Orchestration - Overview

1.5.2 Why Single Agents Hit a Wall

1.5.3 The hub and spok

1.5.4 The Coordinator Role

1.5.5 Subagents

1.5.6 Context Isolation

1.5.7 Coordinator Responsibilities

1.5.8 When to use Multi Agent

1.5.9 Worked Example Research System

1.5.10 Patterns and Takeaways

Knowledge Check: Multi-Agent Orchestration - Hub-and-spoke Architecture

1.6.1 Subagent Spawning - Overview

1.6.2 Subagent as Agentic Loop

1.6.3 Coordinator Orchestrates Subagents

1.6.4 The Agent SDK Approach

1.6.5 Agent Definition Fields

1.6.6 How the Agent Tool works

1.6.7 Context Passing

1.6.8 Processing Subagent Results

1.6.9 Tool Restriction

1.6.10 Patterns and Takeaways

Knowledge Check: Subagent Spawning

1.7.1 Parallel vs Sequential Execution - Overview

1.7.2 Setup the Exercise

1.7.3 Understand the Subagents

1.7.4 Implement Sequential

1.7.5 Spot the Parallelization

1.7.6 Refactor to Parallel Snippet

1.7.7 Handle Partial Failures

1.7.8 Patterns and Takeaways

1.8.1 Multi-Step Workflow Enforcement - Overview

1.8.2 The Autonomy Control Spectrum

1.8.3 Validation Gates

1.8.4 Approval Thresholds

1.8.5 Human in the Loop

1.8.6 Gated Sequences and State Tracking

1.8.7 Prompt vs Programmatic

1.8.8 Handoff Patterns

1.8.9 When You Need Hooks

1.8.10 Patterns and Takeaways

Knowledge Check: Multi-Step Workflow Enforcement and Handoff Patterns

1.9.1 Programmatic Hooks - Overview

1.9.2 The Hook Lifecycle

1.9.3 Blocking Refunds Walkthough

1.9.4 Sandboxing File Writes

1.9.5 Post Tool Use Audit Logging

1.9.6 Configuring Hooks Settings JSON

1.9.7 Hooks vs Coordinator and Lesson Takeaways

Knowledge Check: Programmatic Hooks - PostToolUse and Tool Call Interception

1.10.1 Session Management - Lesson Overview

1.10.2 What is a Session?

1.10.3 Resume Session

1.10.4 Fork Session

1.10.5 Named Sessions

1.10.6 Resume, Fork, or Fresh?

Knowledge Check: Session Management - Resume, Fork, Named Sessions

1.11.1 Domain 1 Scenario Practice

Domain 1 Test

1.12.1 Implementing Workflow Enforcement - Lesson Overview

1.12.2 Set up the Exercise

Welcome to Module 2

Module 2 Pre-Assessment

2.1.1 Tool Descriptions as the Selection Mechanism

2.1.2 How Claude Selects Tools

2.1.3 Tool Definition Schema

2.1.4 Writing Effective Descriptions

2.1.5 Misrouting Anti-Patterns

2.1.6 Consolidation and Namespacing

2.1.7 tool_choice vs Descriptions

2.1.8 Key Takeaways

Knowledge Check: Tool Descriptions as the Selection Mechanism

2.2.1 Writing Effective Tool Interfaces - Overview

2.2.2 Examine the Poor Descriptions

2.2.3 Run the Baseline

2.2.4 Read the Misrouting Report

2.2.5 Apply the Four Question Framework

2.2.6 Test and Iterate

2.2.7 Takeaways and Stretch

2.3.1 Structured Responses - Overview

2.3.2 Error as Contract

2.3.3 Error Category Taxonomy

2.3.4 Error Response Schema

2.3.5 Instructive vs Generic

2.3.6 Multi-Turn Recovery

2.3.7 Common Mistakes

2.3.8 Key Takeaways

Knowledge Check: Structured Error Response

2.4.1 Tool Distribution - Lesson Overview

2.4.2 Least Privilege Framework

2.4.3 Distribution in Practice

2.4.4 Distribution Anti-Patterns

2.4.5 The Four tool_choice modes

2.4.6 Tool Choice and Descriptions

2.4.7 Distribution Plus Tool Choice

2.4.8 Key Takeaways

Knowledge Check: Tool Distribution Across Agents and tool_choice Configuration

2.5.1 MCP Fundamentals - Lesson Overview

2.5.2 What is MCP?

2.5.3 Three Tier Architecture

2.5.4 MCP Primitives

2.5.5 MCP JSON Config

2.5.6 Tool Naming Convention

2.5.7 Complete Flow

2.5.8 Key Takewaways

Knowledge Check: MCP Fundamentals - Servers, Tools, Resources, .mcp.json

2.6.1 Configuring an MCP Server in Claude Code

2.6.2 Create Test Directory

2.6.3 Add MCP Server

2.6.4 Verify and Inspect

2.6.5 Test in Claude Code

2.6.6 Modify Config Manually

2.6.7 Environment Variable Expansion

2.6.8 Troubleshooting and Recap

2.7.1 Built-In Tools: Read, Write, Edit, Glob, Grep, Bash - Lesson Overview

2.7.2 The Six Tools

2.7.3 File Tools

2.7.4 Search Tools

2.7.5 Bash and the Context Trap

2.7.6 Selection Framework

2.7.7 Narrowing Search Pattern

2.7.8 Restrictions and Takeaways

Knowledge Check: Built-In Tools - Read, Write, Edit, Bash, Grep, Glob

2.8.1 Domain 2 Practice Test

2.9.1 Implementing Structured Error Responses

Welcome to Module 3: Claude Code

Domain 3 Module Pre-Assessment

3.1.1 Claude.md Configuration Hierarchy

3.1.2 What is Claude.md?

3.1.3 The Four Levels

3.1.4 Discovery and Merging

3.1.5 Import Syntax

3.1.6 Path-Specific Rules

3.1.7 Choosing the Right Level

3.1.8 Key Takeaways

Knowledge Check: CLAUDE.md Configuration Hierarchy

3.2.1 Building a Claude.md for a Sample Project

3.2.2 Create the Sample Project

3.2.3 Write the Project Claude.md

3.2.4 Add Import References

3.2.5 Create CLAUDE.local.md

3.2.6 Test the Configuration

3.2.7 Refine and Commit

3.2.8 Key Takeaways

3.3.1 Custom Slash Commands and Skills - Lesson Overview

3.3.2 Custom Slash Commands

3.3.3 Skills and Auto Triggering

3.3.4 Context Inherit vs Fork

3.3.5 Allowed Tools and Model

3.3.6 Commands vs Skills

3.3.7 Key Takeaways

Knowledge Check: Custom Slash Commands & Skills (Frontmatter Options)

3.4 Path-Specific Rules with .claude-rules- and Glob Patterns

3.4.1 Path Specific Rules with Claude.md

3.4.2 Create Rules Directory and Unconditional Rule

3.4.3 Path-Scoped Rule

3.4.4 Multi Pattern Rules

3.4.5 Verify Scoping

3.4.6 Organize and Commit

3.4.7 Key Takeaways

3.5.1 Plan Mode vs Direct Execution

3.5.2 What is Plan Mode?

3.5.3 The Descision Framework

3.5.4 Case Study: Simple Bug Fix

3.5.5 Case Study: Database Refactoring

3.5.6 Case Study: Ambiguous Feature

3.5.7 Signals and Key Takeaways

Knowledge Check: Plan Mode vs Direct Execution - Decision Framework

3.6.1 Iterative Refinement, Interview Pattern, Test Driven Iteration, I/O Examples

3.6.2 Interview Pattern

3.6.3 Test Driven Iteration

3.6.4 IO Examples

3.6.5 Choosing the Right Technique

3.6.6 Key Takeaways

Knowledge Check: Iterative Refinement - Interview Pattern, Test-Driven Iteration, I-O Examples

3.7.1 Integrating Claude Code into CI-CD, -p Flag, --output-format --json-schema Overview

3.7.2 -p Flag Non-Interactive

3.7.3 Output format and JSON Schema

3.7.4 Safety Flags

3.7.5 Separate Reviewer Principle

3.7.6 CI-CD Pipeline Example

3.7.7 Key Takeaways

Knowledge Check: Integrating Claude Code into CI-CD - -p Flag, --output-format, --json-schema

3.8.1 The Explore Subagent compact and Session Management Overview

3.8.2 Explore Subagent

3.8.3 Compact Command

3.8.4 Session Management

Knowledge Check: The Explore Subagent, compact, and Session Management

3.9.1 Domain 3 Practice Test

Domain 3 Test

Module 4 Overview: Prompt Engineering

Domain 4 Pre-Assessment

4.1.1 Explicit Criteria vs Vague Instructions

4.1.2 The Interpretation Gap

4.1.3 Vague vs Explicit

4.1.4 The Critera Framework

Knowledge Check: Explicit Criteria vs Vague Instructions

4.2.1 Few-Shot Prompting for Consistency and Ambiguity Handling - Overview

4.2.2 Why Few-Shot Works

4.2.3 Structuring Examples

4.2.4 Choosing Examples

4.2.5 When to Use Few-Shot

Knowledge Check: Few-Shot Prompting for Consistency and Ambiguity Handling

4.3.1 Few Shot Practical Examples

4.3.2 Run the Baseline

4.3.3 Examine Test Messages

4.3.4 Write Few-Shot Examples

4.3.5 Test and Iterate

4.4.1 Structured Output Via Tool Use

4.4.2 The Pattern

4.4.3 Schema Design

4.4.4 Other and Full Schema

4.4.5 When to Use

Knowledge Check: Structured Output via tool_use - JSON Schemas, Enums, Nullable Fields

4.5.1 Schema Design Patterns

4.5.2 Run the Basic Schema

4.5.3 Examine the Documents

4.5.4 Design the Improved Schema

4.5.5 Test and Compare

4.6.1 Retry Loops

4.6.2 The Validation Retry Pattern

4.6.3 Writing Error Feedback

4.6.4 Implementing the Loop

4.6.5 Limits and Routing

Knowledge Check: Validation-Retry Loops and Feedback-Driven Self-Correction

4.7.1 Batch Processing

4.7.2 What the Batch API is

4.7.3 The Batch Lifecycle

4.7.4 Retrieving and Matching Results

4.7.5 When to Use Batch

Knowledge Check: Batch Processing with the Message Batches API

4.8.1 Multi-Pass Review Architecture and Prompt Chaining

4.8.2 Why Multi Pass

4.8.3 Generate Review Pattern

4.8.4 Coarse-Fine Pattern

4.8.5 When Multi-Pass is Worth It

Knowledge Check: Multi-Pass Review Architectures and Prompt Chaining

4.9.1 Domain 4 Practice Exam

Domain 4 Practice Exam

Module 5 Introduction

Module 5 Pre-Assessment

5.1.1 Context Fundamentals - Overview

5.1.2 What Consumes Tokens

5.1.3 Token Counting Heuristics

5.1.4 The Lost-in-the-Middle Effect

5.1.5 Mitigation Strategies

5.1.6 Tool Output Accumulated

5.1.7 Preserving Info in Long Sessions

5.1.8 Key Takeaways

Knowledge Check: Context Window Fundamentals - Token Budgets, Lost-in-the-Middle Effect

5.2.1 Trimming Tool Outputs and Structured Fact Extraction

5.2.2 Measuring the Problem

5.2.3 Tool Level Trimming

5.2.4 Orchestrator Level Trimming

5.2.5 Structured Subagent Returns

Knowledge Check: Trimming Tool Outputs and Structured Fact Extraction

5.3.1 Progressive Summarization Risks and Case Facts Blocks

5.3.2 Four Risks

5.3.3 Case Facts Block

5.3.4 Multi Issue Sessions

5.3.5 Implementing the Pattern

Knowledge Check: Progressive Summarization Risks and Case Facts Blocks

5.4.1 Escalation Patterns

5.4.2 The Three Triggers

5.4.3 Explicit Requests

5.4.4 Policy Gap

5.4.5 Frustrated Customer

5.4.6 Ambiguous Matches

5.4.7 Key Takeaways

Knowledge Check: Escalation Patterns - When to Escalate vs Resolve

5.5.1 Error Propagation in Multi-Agent Systems

5.5.2 Why Error Propagation Matters

5.5.3 Structured Error Context

5.5.4 Access Failure vs Empty Result

5.5.5 Local Recovery and Anti-Patterns

5.5.6 Coverage Annotations

5.5.7 Key Takeaways

Knowledge Check: Error Propagation in Multi-Agent Systems

5.6.1 Scratchpad Recovery

5.6.2 Context Degradation Symptoms

5.6.3 Scratchpad Pattern

5.6.4 Progressive Discovery Example

5.6.5 Subagent Delegation and Phases

5.6.6 Crash Recovery State Manifest

5.6.7 Key Takeaways

Knowledge Check: Scratchpad Files, Crash Recovery, and Structured State Persistence

5.7.1 Human Review Workflows - Confidence, Calibration and Stratification

5.7.2 Aggregate Accuracy Trap

5.7.3 Stratified vs Random Sampling

5.7.4 Field Level Confidence

5.7.5 Calibrating Thresholds

5.7.6 Routing and Feedback Loop

5.7.7 Key Takeaways

Knowledge Check: Human Review Workflows - Confidence Calibration and Stratified Sampling

5.8.1 Lesson Overview

Domain 5 Practice Test

Introduction to Practical Capstones

6.1.1 Capstone A - Overview

6.1.2 Set up the Exercise

6.1.3 Implement the Agent Loop

6.1.4 Structured Error Responses

6.1.5 Refund Threshold Hook

6.1.6 Escalation Logic

6.1.7 Context Management

6.1.8 Validate and Recap

6.2.1 - Capstone Project B - Overview

6.2.2 Set up the Exercise

6.2.3 Project Claude.md

6.2.4 Path Scoped Roles

6.2.5 Command and Skill

6.2.6 MCP Server

6.2.7 Self Assessment and Recap

6.3.1 - Capstone Project C

6.3.2 Set up and Review Documents

6.3.3 Design the Schema

6.3.4 Build the Extractor

6.3.5 Validation and Retry

6.3.6 Human Review Routing

6.3.7 Run the Full Pipeline

6.3.8 Validate and Recap

7.1.1 Preparing for the Exam

7.1.2 Exam at a Glance

7.1.3 Six Scenarios: Part 1

7.1.4 Six Scenarios Part 2

7.1.5 Three Layers of a Question

7.1.6 Distractor Patterns

7.1.7 Worked Example: Customer Support

7.1.8 Worked Example: Claude Code

7.1.9 Worked Example Data Extraction

7.1.10 Elimination Framework

7.1.11 Common Mistakes

Knowledge Check: Anatomy Of A Scenario-Based Exam Question

7.2.1 The Practice Exam

Full Length Practice Exam

7.3.1 Practice Exam Post-Mortem