CertSafari

    Anthropic Claude Certified Architect – Foundations (CCAR-F) Lessons

    30 lessons, one per exam-guide subdomain, in the order the guide teaches them. Every claim is cited to the official documentation.

    Domain 1: Agentic Architecture & Orchestration

    7 lessons · 27% of the exam

    1. 1.1Agentic loop lifecycle: stop_reason control flow and tool results

      Subdomain 1.1: Design and implement agentic loops for autonomous task execution

      2 pages · 22 min read

    2. 1.2Hub-and-Spoke Coordinators: Context Isolation and Routing

      Subdomain 1.2: Orchestrate multi-agent systems with coordinator-subagent patterns

      2 pages · 23 min read

    3. 1.3Spawning Subagents: the Agent (Task) Tool and AgentDefinition

      Subdomain 1.3: Configure subagent invocation, context passing, and spawning

      2 pages · 19 min read

    4. 1.4Prerequisite Gates: Enforcing Workflow Order with Hooks

      Subdomain 1.4: Implement multi-step workflows with enforcement and handoff patterns

      2 pages · 21 min read

    5. 1.5Agent SDK Hooks: Intercept Tool Calls and Normalize Results

      Subdomain 1.5: Apply Agent SDK hooks for tool call interception and data normalization

      16 min read

    6. 1.6Prompt Chaining vs Dynamic Decomposition

      Subdomain 1.6: Design task decomposition strategies for complex workflows

      2 pages · 19 min read

    7. 1.7Resuming Claude Sessions by Name and ID

      Subdomain 1.7: Manage session state, resumption, and forking

      2 pages · 23 min read

    Domain 2: Tool Design & MCP Integration

    5 lessons · 18% of the exam

    1. 2.1Tool Descriptions: How Claude Chooses Which Tool to Call

      Subdomain 2.1: Design effective tool interfaces with clear descriptions and boundaries

      2 pages · 23 min read

    2. 2.2MCP Tool Errors: isError vs Protocol Errors

      Subdomain 2.2: Implement structured error responses for MCP tools

      2 pages · 20 min read

    3. 2.3Scoped Tool Access for Multi-Agent Systems

      Subdomain 2.3: Distribute tools appropriately across agents and configure tool choice

      2 pages · 17 min read

    4. 2.4MCP server scopes and secrets in Claude Code

      Subdomain 2.4: Integrate MCP servers into Claude Code and agent workflows

      2 pages · 20 min read

    5. 2.5Grep vs Glob vs Read: Searching a Codebase with Claude Code's Built-in Tools

      Subdomain 2.5: Select and apply built-in tools (Read, Write, Edit, Bash, Grep, Glob) effectively

      2 pages · 18 min read

    Domain 3: Claude Code Configuration & Workflows

    6 lessons · 20% of the exam

    1. 3.1CLAUDE.md Hierarchy: User, Project and Directory Scopes

      Subdomain 3.1: Configure CLAUDE.md files with appropriate hierarchy, scoping, and modular organization

      2 pages · 20 min read

    2. 3.2Slash Commands and Skills: Project vs Personal Scope

      Subdomain 3.2: Create and configure custom slash commands and skills

      2 pages · 23 min read

    3. 3.3Path-Scoped Rules in .claude/rules/: paths Frontmatter and Glob Patterns

      Subdomain 3.3: Apply path-specific rules for conditional convention loading

      2 pages · 18 min read

    4. 3.4Plan Mode vs Direct Execution: Choosing by Task Scope

      Subdomain 3.4: Determine when to use plan mode vs direct execution

      2 pages · 18 min read

    5. 3.5Examples, Tests First and the Interview Pattern

      Subdomain 3.5: Apply iterative refinement techniques for progressive improvement

      2 pages · 16 min read

    6. 3.6Headless Claude Code in CI: -p, Tool Permissions and JSON Output

      Subdomain 3.6: Integrate Claude Code into CI/CD pipelines

      2 pages · 22 min read

    Domain 4: Prompt Engineering & Structured Output

    6 lessons · 20% of the exam

    1. 4.1Explicit Review Criteria vs Vague Instructions in Claude Prompts

      Subdomain 4.1: Design prompts with explicit criteria to improve precision and reduce false positives

      2 pages · 19 min read

    2. 4.2Few-Shot Prompting for Consistent Output Format

      Subdomain 4.2: Apply few-shot prompting to improve output consistency and quality

      2 pages · 21 min read

    3. 4.3Extraction Tools and tool_choice: Forcing Structured Output

      Subdomain 4.3: Enforce structured output using tool use and JSON schemas

      2 pages · 20 min read

    4. 4.4Retry with Error Feedback in Extraction Pipelines

      Subdomain 4.4: Implement validation, retry, and feedback loops for extraction quality

      2 pages · 20 min read

    5. 4.5Message Batches API: When to Batch and When to Stay Synchronous

      Subdomain 4.5: Design efficient batch processing strategies

      2 pages · 20 min read

    6. 4.6Independent Review Instances vs. Self-Review

      Subdomain 4.6: Design multi-instance and multi-pass review architectures

      2 pages · 22 min read

    Domain 5: Context Management & Reliability

    6 lessons · 15% of the exam

    1. 5.1Keeping Critical Facts Intact in Long Conversations

      Subdomain 5.1: Manage conversation context to preserve critical information across long interactions

      2 pages · 24 min read

    2. 5.2Escalation Triggers and Ambiguity Resolution for Support Agents

      Subdomain 5.2: Design effective escalation and ambiguity resolution patterns

      16 min read

    3. 5.3Error Propagation in Multi-Agent Systems

      Subdomain 5.3: Implement error propagation strategies across multi-agent systems

      14 min read

    4. 5.4Context Rot, /compact and Scratchpad Files in Long Codebase Sessions

      Subdomain 5.4: Manage context effectively in large codebase exploration

      2 pages · 19 min read

    5. 5.5Human Review Routing and Confidence Calibration for Extraction

      Subdomain 5.5: Design human review workflows and confidence calibration

      13 min read

    6. 5.6Claim-Source Mappings: Keeping Provenance Through Multi-Agent Synthesis

      Subdomain 5.6: Preserve information provenance and handle uncertainty in multi-source synthesis

      2 pages · 16 min read

    Practice Anthropic Claude Certified Architect – Foundations (CCAR-F) in quiz mode