Claude Code Certified Professional — Claude Cert Academy

Master the tool that codes with you.

A standalone certification covering Claude Code end-to-end: from first install through Agent SDK mastery. Seven modules in linear progression (CLI, configuration, surfaces, hooks/skills/plugins, MCP, CI/CD, Agent SDK) with scenario-based practice questions per module and a practical capstone project. Designed for developers who want to go beyond casual use into power-user and platform-builder territory.

What you'll learn

• Install, configure, and operate Claude Code across CLI, IDE, Desktop, and Web surfaces
• Design CLAUDE.md instruction hierarchies and auto-memory strategies for teams
• Build custom hooks, skills, and plugins that automate development workflows
• Connect external tools and data sources via MCP server configuration
• Integrate Claude Code into CI/CD pipelines using GitHub Actions and GitLab CI
• Build custom agents with the Agent SDK: streaming, subagents, structured output, observability
• Administer Claude Code at organizational scale: permissions, cost controls, analytics

Course content

01. Getting Started

• How Claude Code works: the agentic loop, tools, and context window
• Installing Claude Code: macOS, Linux, Windows, Homebrew, WinGet
• Authentication: Anthropic, Amazon Bedrock, Google Vertex AI, Microsoft Foundry
• Your first real task: from problem to commit
• The CLI reference: flags, headless mode, piping, and shell scripting
• Interactive mode: keyboard shortcuts, slash commands, and input methods
• Permission modes: default, auto, plan — and how the sandbox works
• Troubleshooting: the 12 most common setup failures and how to fix them

02. Configuration & Memory

• The CLAUDE.md hierarchy: global, project, and local scopes
• Writing great CLAUDE.md files: rules, imports, and anti-patterns
• Auto-memory: what Claude saves, where it lives, how to control it
• Settings files: settings.json, settings.local.json, server-managed settings
• Environment variables: API keys, model overrides, and custom config
• The .claude directory: structure, keybindings, project conventions
• Model configuration: selecting models, fast mode, and thinking modes
• Context window management: what loads, what compresses, and how to optimize
• Team CLAUDE.md strategies: shared rules, per-engineer overrides, onboarding

03. Surfaces & Integrations

• VS Code extension: inline diffs, @-mentions, conversation history
• JetBrains plugin: full Claude Code engine in IntelliJ, PyCharm, WebStorm
• Claude Code Desktop: parallel sessions and Dispatch
• Claude Code on the Web: browser sessions and --teleport
• Slack integration: @Claude in channels and bug-to-PR workflow
• Remote Control: moving sessions between surfaces without losing context
• Voice dictation: /voice command and spoken prompt patterns
• Chrome extension (beta): live DOM, network, and console access

04. Hooks, Skills & Plugins

• Hooks: the event model, shell execution, and lifecycle
• PreToolUse and PostToolUse hooks: intercept, transform, block
• Notification and Stop hooks: observability and clean exits
• Hook security: injection risks, input validation, sandboxing
• Build: an auto-format hook that runs on every file edit
• Skills: frontmatter schema, tool access declarations, and scope
• Writing a custom skill from scratch
• Plugins: package structure, distribution, and marketplaces
• Skills vs. plugins: when to use each and how they compose
• Sub-agents: spawning, frontmatter types, and agent teams

05. MCP & Tool Connections

• Model Context Protocol: why it exists, transport types
• Configuring MCP servers: scope levels, auth, and the settings schema
• Pre-built integrations: Google Drive, Jira, Slack, GitHub
• Writing your own MCP server: schema, sampling, and notifications
• Tool search: scaling to hundreds of tools without context bloat
• MCP security: authorization, dangerous tool patterns, least-privilege
• Debugging MCP connections: common errors and validation strategies
• MCP in production: versioning, monitoring, and fallback patterns

06. CI/CD & Scheduling

• Headless mode: running Claude Code non-interactively in scripts
• GitHub Actions: setup, secrets, triggers, and the claude-code action
• Automated PR review with GitHub Actions
• GitLab CI/CD: runner setup and pipeline integration
• Routines: Anthropic-managed scheduling, webhook triggers, and events
• Desktop scheduled tasks: local schedule with full file access
• /loop: session-scoped recurring prompts and self-pacing
• Cost tracking and usage analytics in CI contexts

07. Agent SDK

• Agent SDK overview: SDK vs. CLI vs. API
• Quickstart: Python and TypeScript setup, your first agent
• The agent loop: turns, stop reasons, and tool execution
• Sessions: creating, resuming, and managing state
• Streaming output: real-time responses and partial block handling
• Structured output: JSON schema enforcement and validation retry loops
• Custom tools: writing schemas and handler functions
• Hooks in the SDK: programmatic PreToolUse and PostToolUse control
• Subagents in the SDK: spawning, coordination, and result collection
• Permissions and sandboxing in SDK agents
• Observability: OpenTelemetry, cost tracking, and structured logging
• Hosting and secure deployment: production patterns for Agent SDK apps

Continue to Claude Cert Academy