Skip to content

Amplihack Tutorial

Comprehensive guide to mastering amplihack from basics to advanced topics

Overview

This tutorial provides a complete learning path through amplihack's capabilities, from your first workflow to building custom goal-seeking agents. Designed for 60-90 minutes of hands-on learning with progressive disclosure based on your skill level.

How to Start the Tutorial

The simplest way — just type this in Claude Code:

"I want to learn amplihack - take me through the tutorial"

Claude will automatically invoke the guide agent and begin the tutorial.

What's a "guide agent"? Amplihack includes specialized AI agents for different tasks (building, reviewing, testing, etc.). The guide agent is one that's designed specifically for interactive teaching. You don't need to understand agents to use one — Claude handles the delegation.

For advanced users — you can invoke the guide agent directly using Claude Code's Task() tool (a built-in function that delegates work to specialized agents):

Task(subagent_type="guide", prompt="Start tutorial")

The guide agent will:

  1. Assess your skill level
  2. Tailor explanations to your experience
  3. Provide platform-specific examples
  4. Walk you through 7 comprehensive sections

What You'll Learn

Section 1: Welcome & Setup (5 min)

  • What amplihack is and why it exists
  • Core philosophy: ruthless simplicity, modular design, zero-BS
  • Quick environment check
  • Navigation tips

Section 2: First Workflow (10 min)

  • Execute your first amplihack workflow
  • Understand the 22-step default workflow
  • Watch agents collaborate
  • Verify successful completion

Section 3: Workflows Deep Dive (15 min)

  • All 8 workflows and when to use each
  • Workflow selection framework
  • Q&A, Default, Investigation, Auto modes
  • Fault-tolerant workflows (N-version, Debate, Cascade)
  • Document-Driven Development (DDD)

Section 4: Prompting Techniques (15 min)

  • Anatomy of a great prompt
  • Structuring complex requests
  • Leveraging context and constraints
  • Platform-specific prompting patterns
  • Common mistakes and how to avoid them

Section 5: Continuous Work (15 min)

  • Auto mode for multi-turn autonomous execution
  • Lock mode for keeping agents working
  • Injecting instructions mid-session
  • Monitoring and controlling autonomous agents

Section 6: Goal Agents (15 min)

  • What goal-seeking agents are
  • Creating agents from simple prompts
  • Running and monitoring goal agents
  • When to use goal agents vs workflows

Section 7: Advanced Topics (15 min)

  • Skills system (74+ capabilities)
  • Hooks for customizing behavior
  • Memory systems (5-type, Neo4j)
  • Power features and customization

Learning Paths

The tutorial adapts to your experience level:

Beginner Path (90 minutes):

Section 1 → Section 2 → Section 3 → Section 4
└─ Focus: Understanding basics and running your first workflow

Intermediate Path (60 minutes):

Section 2 → Section 3 → Section 5
└─ Focus: Workflows and autonomous execution

Advanced Path (60 minutes):

Section 3 → Section 6 → Section 7
└─ Focus: Advanced workflows, goal agents, and customization

Features

Progressive Disclosure

Content tagged by skill level:

  • [BEGINNER] - Extra explanation for new users
  • [INTERMEDIATE] - Practical application details
  • [ADVANCED] - Deep technical details and customization

Platform Support

Examples for all supported platforms:

  • Claude Code
  • Amplifier
  • GitHub Copilot CLI
  • OpenAI Codex
  • RustyClawd

Interactive Navigation

Jump to any section:

"Section 3"     - Jump to Workflows Deep Dive
"Continue"      - Go to next section
"Menu"          - Show all sections
"Section 5"     - Jump to Continuous Work

Hands-On Exercises

"Try It Now" sections with real examples:

  • Execute your first workflow
  • Create a goal-seeking agent
  • Use auto mode
  • Customize preferences

Expected Outcomes

After completing this tutorial, you'll be able to:

✅ Execute the default 22-step workflow ✅ Choose the right workflow for any task ✅ Write effective prompts for AI agents ✅ Use auto mode for autonomous multi-turn work ✅ Create custom goal-seeking agents ✅ Customize amplihack for your needs ✅ Understand skills, hooks, and memory systems

Troubleshooting

Tutorial Won't Start

Issue: Guide agent not responding

Solution:

# Ensure you're in amplihack environment
amplihack --version

# Try explicit agent invocation
Task(subagent_type="guide", prompt="Start tutorial from Section 1")

Issue: Jumping to sections doesn't work

Solution: Use explicit section names in quotes:

"Take me to Section 3: Workflows Deep Dive"

Can't Complete Exercises

Issue: Hands-on exercises fail

Solution:

  • Check Prerequisites
  • Verify API keys set
  • Ensure Git and platform CLIs installed

Getting Help

During Tutorial:

  • Type "Help" for assistance
  • Ask specific questions at any time
  • Request clarification on any topic

After Tutorial:

Feedback

Help us improve this tutorial:

  1. Report issues or unclear sections
  2. Suggest additional topics
  3. Share your learning path preferences

Create an issue: Tutorial Feedback

Ready to start? Invoke the guide agent and begin your amplihack journey!

"Start the amplihack tutorial"