Claude Code: An agentic adventure

A hands-on workshop where you learn to ship real software using AI coding agents — no prior programming experience required.

Description

You will build a working web application from scratch using a team of AI agents. Instead of writing code yourself, you direct agents through the full cycle: specifying what to build, assembling a specialized agent team, launching them on your project, and deploying the result. By the end of the workshop, you walk out with a deployed application.

The workshop follows a structured flow across six modules:

1. Launch Pad — Install Claude Code and set up your development environment
2. Mission Briefing — Understand how AI agents work, why they change the game, and what you'll build
3. Spec Your Payload — Create an impact map, write a feature spec, and scaffold your project
4. Assemble the Crew — Plan, build, and assemble a multi-agent team with defined roles
5. Launch — Run your agents, iterate and debug, and get to working software
6. Splashdown — Reflect on what you built and look ahead at what's next

This workshop can scale to large groups if need be.

Learning Goals

• Explain what AI agents are and how the agent loop works [Understand]
• Evaluate which tasks to delegate to agents versus handle yourself [Evaluate]
• Employ an agent team to map impact and specify features [Apply]
• Operate a multi-agent team with defined roles [Apply]

Subjects

• What AI agents are and how the agent loop works (observe, act, learn, repeat)
• Three layers of instructions: system, project, and task
• Why multiple specialist agents produce better results than a single agent
• The 4D framework: Delegate, Describe, Discern, Diligence
• Impact mapping to define project scope and MVP features
• Writing specifications that guide agent teams effectively
• Planning and building a multi-agent team with defined roles
• Launching agents, iterating, and debugging to reach working software
• Deploying your application