Skip to content

syllabus.md

Latest commit

 

History

History
156 lines (132 loc) · 8.81 KB

syllabus.md

File metadata and controls

156 lines (132 loc) · 8.81 KB
Raw

Claude Code: Zero To Hero

For Developers, DevOps, Cloud, SRE, and Platform Engineers — Anyone Who Wants to Actually Use Claude Code

Format: Self-paced, recorded | Language: English | Channel: trainwithshubham.ai Built from: Anthropic's official Claude Code documentation and changelog as of August 2026

How This Course Is Structured

16 modules across 4 phases. Everyone takes Modules 1–14. Module 13 is a DevOps/Cloud deep-dive every learner should still watch (even the "general dev" track — it's what makes this course different from a generic AI coding tutorial). Module 15 branches into two capstone tracks so a frontend/backend developer and an SRE both leave with a portfolio project that matches their actual job.

Phase Modules What it covers
Zero → Comfortable 1–3 AI/agent fundamentals, prompt engineering for agents, installing Claude Code everywhere, your first session
Comfortable → Productive 4–9 The agentic loop and built-in tools, permissions & config, planning real projects, Skills/Hooks/Plugins, MCP
Productive → Advanced 10–13 Multi-agent orchestration, every surface (Desktop/Web/Mobile/Slack), CI/CD & security, enterprise/cloud deployment
Advanced → Hero 14–16 Agent SDK, capstone (choose your track), staying current

Module 1 — Welcome to the Era of Agentic Coding

  1. What is Generative AI and a Large Language Model (quick refresher)
  2. Claude's Model Lineup in 2026: Haiku 4.5, Sonnet 5, Opus 5, and the Mythos Tier
  3. Claude.ai vs Claude API vs Claude Code vs Claude Cowork vs Agent SDK — Choosing the Right Surface
  4. What Makes an AI System "Agentic": The Agentic Loop Explained
  5. Where Claude Code Fits — For Developers, DevOps, SRE, Cloud and Platform Engineers
  6. Course Roadmap: How to Use This Course and the Two Capstone Tracks

Module 2 — Prompt Engineering for Agentic Work

  1. Anatomy of a Good Prompt for an Agent (not a chatbot)
  2. System Prompts, CLAUDE.md, and Output Styles — Where Instructions Actually Live
  3. Zero-Shot vs Few-Shot vs Chain-of-Thought for Coding Tasks
  4. Common Mistakes When Prompting an Autonomous Agent
  5. Hands-On: Writing Your First Effective Prompts in Claude Code

Module 3 — Installing Claude Code Everywhere

  1. Installing on macOS, Windows, and Linux (native binaries)
  2. Full Platform Tour: CLI vs VS Code vs JetBrains vs Desktop vs Web vs Mobile
  3. Authentication and Account Setup — Individual, Team, Enterprise
  4. Your First Claude Code Session
  5. Running Claude Code on Other Providers — Bedrock, Google Cloud, Foundry, and Gateways
  6. Interface Deep Dive: Interactive Mode, Keyboard Shortcuts, Themes

Module 4 — The Agentic Loop and Built-In Tools

  1. How Claude Code Works: Reasoning, Tool Use, Iteration
  2. Read, Write, Edit — File Operation Tools
  3. Grep and Glob — Searching Large Codebases
  4. The Bash Tool and the Sandboxed Bash Tool
  5. Understanding the Context Window and Prompt Caching
  6. Tools Reference Walkthrough

Module 5 — Permissions, Memory, and Configuration

  1. Choosing a Permission Mode: Manual, Accept Edits, Auto Mode
  2. Auto Mode Deep Dive — The New Default and How the Classifier Decides
  3. CLAUDE.md Explained: User, Project, and Directory Levels
  4. Auto Memory — How Claude Remembers Your Project Over Time
  5. Exploring the .claude Directory and settings.json
  6. Debugging Your Config: /context, /doctor, /hooks, /mcp

Module 6 — Planning and Executing a Real Project

  1. Green Field vs Brown Field Projects
  2. Plan Mode vs Direct Execution vs Auto Mode — When to Use Which
  3. Session Management: Resume, Branch, Rewind, Checkpointing
  4. Working Across a Monorepo or Large Codebase
  5. Common Workflows: Bug Fixes, Refactors, Test Generation
  6. Hands-On: Planning and Building a Starter Project

Module 7 — Extending Claude Code: Skills, Commands, and Hooks

  1. Introduction to Agent Skills (SKILL.md)
  2. Creating Custom Slash Commands
  3. Automating Actions with Hooks (formatting, notifications, validation)
  4. Output Styles — Adapting Claude Code Beyond Software Engineering
  5. Hands-On: Building Your Own Skill

Module 8 — The Plugin Ecosystem

  1. What Are Plugins and Why They Matter
  2. Discovering and Installing Plugins from Marketplaces
  3. Creating and Distributing Your Own Plugin Marketplace
  4. Claude Security Plugin — Scanning Your Codebase for Vulnerabilities
  5. security-guidance Plugin — Catching Issues as Claude Writes Code
  6. Hands-On: Installing and Configuring a Plugin Stack

Module 9 — MCP (Model Context Protocol) in 2026

  1. What Is MCP and Why It Became the Industry Standard
  2. MCP Architecture: The 2026-07-28 Spec, Stateless Core, OAuth/OIDC
  3. Connecting Your First MCP Server (claude mcp login)
  4. Real-World MCP Example: AWS and Observability Tools
  5. Channels — Pushing Alerts and Webhooks Into a Running Session
  6. Hands-On: Wiring an MCP Server Into Your Own Stack

Module 10 — Subagents and Orchestration at Scale

  1. Creating Custom Subagents for Task-Specific Work
  2. Background Subagents and Agent View — Managing Many Sessions at Once
  3. Agent Teams — Coordinating Multiple Claude Code Instances
  4. Dynamic Workflows — Orchestrating Dozens to Hundreds of Subagents
  5. Worktrees — Running Parallel Sessions Without Collisions
  6. Cross-Session Messaging — Let Sessions Talk to Each Other
  7. Hands-On: Building a Multi-Agent Codebase Audit

Module 11 — Claude Code Across Every Surface

  1. Your IDE — VS Code and JetBrains
  2. Claude Code Desktop: Parallel Sessions, Computer Use, In-App Browser
  3. Claude Code on the Web and Routines (Scheduled Cloud Agents)
  4. Claude Code on Mobile — Monitor and Steer from Your Phone
  5. Remote Control — Continuing a Session From Any Device
  6. Claude Tag — Bringing Claude Into Slack
  7. Claude in Chrome — Testing and Debugging Web Apps

Module 12 — CI/CD, Code Review, and Security

  1. Headless Mode — Claude Code Without a Terminal (claude -p, --bare, JSON output)
  2. Claude Code GitHub Actions — @claude Mentions and Auto-Fix PRs
  3. Claude Code in GitLab CI/CD
  4. Automated Code Review and /ultrareview — Multi-Agent Bug Hunting
  5. Security Fundamentals: Sandbox Environments and Threat Models
  6. Managing Costs: Token Usage, Model Selection, Effort Levels
  7. Hands-On: Wiring Claude Code Into a CI/CD Pipeline

Module 13 — Claude Code for Infrastructure, Cloud, and Enterprise

(Watch this even if you're doing the general-dev capstone — this is the module that makes "Zero to Hero" different from every other Claude Code tutorial.)

  1. Claude Code for Infrastructure as Code: Terraform, Kubernetes, Docker
  2. Deploying on Amazon Bedrock, Google Cloud's Agent Platform, and Microsoft Foundry
  3. Self-Hosted Environments — Running Cloud Sessions on Your Own Infrastructure
  4. Claude Apps Gateway — Centralized Credentials, Spend Limits, SSO
  5. Admin Setup — Managed Settings, MCP Allowlists, Policy Enforcement
  6. Observability — OpenTelemetry Monitoring and the Analytics Dashboard

Module 14 — Building Products with the Agent SDK

  1. Agent SDK Overview: Claude Code as a Library
  2. The Agent Loop — Message Lifecycle and Architecture
  3. Giving Claude Custom Tools and Structured Outputs
  4. Subagents, Skills, and Plugins Inside the SDK
  5. Hosting the Agent SDK in Production (Docker/Kubernetes)
  6. Tie-In: Pairing the Agent SDK with Durable Execution (Temporal)

Module 15 — Capstone: Choose Your Track

  1. Capstone Kickoff — Planning with Plan Mode + Auto Mode
  2. Track A (General Dev): Ship a Full-Stack SaaS Feature End-to-End with Claude Code
  3. Track B (DevOps/SRE): Build a Self-Healing CI/CD or Incident-Response Agent
  4. Build Sprint (track-specific, guided)
  5. Capstone Review with /code-review and /ultrareview
  6. Presenting and Documenting Your Project with Artifacts

Module 16 — Staying Sharp: What's Next

  1. Reading the Weekly Changelog — How to Stay Current With a Fast-Moving Tool
  2. Claude's Expanding Surfaces — Where This Is Headed
  3. Building Your Own Learning Loop and Community
  4. Course Recap, Certification, and Next Steps

Notes for Production

  • Every module references named, current features (auto mode, agent teams, dynamic workflows, MCP 2026-07-28, Claude Security plugin, self-hosted environments, etc.) pulled from Anthropic's official docs/changelog as of Aug 12, 2026 — not third-party recap blogs, which had inconsistent claims.
  • Claude Code ships weekly, so a "last verified" note in each module's description (like your existing course does) is worth keeping.
  • Module 13 and 14 are the modules that differentiate this from generic "vibe coding" tutorials — worth extra production polish given your DevOps/Cloud/SRE audience overlap with Temporal/AIOps India.
  • Two-track capstone (Module 15) means you'll want two guided project repos/briefs, similar to how you built the DevBoard Live hackathon brief for Udaan Batch 11.