Learning Lab

Master production agent workflows with empirical verification patterns

Safe sandbox for learning and building phased workflow skills with strict validation. Move beyond "LLM writes text" to "LLM executes validated procedures with evidence."

8 Learning Guides
11 Jupyter Notebooks

What You'll Learn

Test Harness Pattern

Understand the mental model behind evidence-based LLM workflows

Multi-Phase Orchestration

Build workflows with strict validation gates between phases

Empirical Verification

Use scripts to validate LLM outputs instead of trusting them

Production Deployment

Take your workflows from learning lab to production systems

Guided Learning Paths

Beginner

1 hour

Get started with the test harness pattern

Advanced

3 hours

Deep dive into orchestration patterns

Reference

As needed

Visual maps and reference documentation

Interactive Jupyter Notebooks

Launch instantly in Google Colab - no local setup required

Agent Skills (5 notebooks)

What is SKILL.md?

Open in Colab View on GitHub

Skill Anatomy

Open in Colab View on GitHub

Build Your First Skill

Open in Colab View on GitHub

Advanced Skill Patterns

Open in Colab View on GitHub

Skill Validation

Open in Colab View on GitHub

Plugin Development (4 notebooks)

What is a Plugin?

Open in Colab View on GitHub

Plugin Structure

Open in Colab View on GitHub

Build Your First Plugin

Open in Colab View on GitHub

MCP Server Plugins

Open in Colab View on GitHub

Orchestration Patterns (2 notebooks)

Mental Model

Open in Colab View on GitHub

Orchestration Pattern

Open in Colab View on GitHub

Getting Started

01

Read the Introduction

Start with GUIDE 00: Start Here to understand the mental model (5 min)

02

Choose Your Path

Pick Beginner, Advanced, or Reference based on your experience level

03

Try Jupyter Notebooks

Launch interactive tutorials in Google Colab - no setup required

04

Build Your Own

Adapt the reference implementation for your production use case