Agentic AI in Platform Engineering

A PlatformEngineering.org course

This repository contains the hands-on exercise code for every module of the course. Each moduleN/ directory is self-contained: one agent script, one sample data file, one config, and a README with the exercise brief.

---

Quick Start

# 1. Fork the repo first (required for GitHub Actions secrets)# Go to https://github.com/InternalDeveloperPlatform/pe-agentic-course and click Fork# 2. Clone YOUR fork
git clone https://github.com/YOUR_USERNAME/pe-agentic-course.git
cd pe-agentic-course
# 2. Set your API keyexport ANTHROPIC_API_KEY=sk-ant-...
# 3. Verify your environment
python shared/verify_setup.py
# 4. Run a module exercise (example: Module 2)
python module2/agent.py

---

Repository Structure

.
├── shared/ # Common utilities — do not modify
│ ├── claude_client.py # ask() function — wraps Anthropic API
│ ├── output.py # JSON, Step Summary, and GitHub Issue formatters
│ └── verify_setup.py # Pre-flight environment check
│
├── module1/ # Module 1: Platform Pain Points & The AI Opportunity
├── module2/ # Module 2: Build Your First AI Agent
├── module3/ # Module 3: Agents That Think — ReAct & Planning
├── module4/ # Module 4: AI-Powered Diagnosis and Remediation
├── module5/ # Module 5: Intelligent CI/CD and Adaptive Delivery
├── module6/ # Module 6: Operational Intelligence & Conversational Observability
├── module7/ # Module 7: Multi-Agent Coordination & Implementation Strategy
├── module8/ # Module 8: Capstone — Build Your Platform Engineering Agent
│
├── .github/workflows/ # One CI workflow per module
│ ├── module1-hello-agent.yml
│ ├── module2-first-agent.yml
│ └── ...
│
├── output/ # Agent output JSON from previous runs (for comparison)
└── docs/ # Architecture diagrams and reference guides

Files in Every Module Directory

File	Purpose
agent.py or triage_agent.py	Main agent script — the exercise entry point
sample_log.txt or sample_data.json	Realistic test data for the exercise
agent-config.yml	Scenario config: model, max_iterations, context fields
README.md	Exercise brief, setup instructions, success criteria

---

Prerequisites

Requirement	Version	Notes
Python	≥ 3.10
anthropic SDK	latest	pip install anthropic
GitHub CLI (gh)	any	Optional — needed for Modules 7–8
ANTHROPIC_API_KEY	—	Set as env var or GitHub Secret

---

GitHub Actions

Each module has a corresponding workflow in .github/workflows/. Workflows trigger on push to the relevant moduleN/** path and run the agent against the sample data. Output JSON is uploaded as an artifact for comparison.

To enable CI for your fork: add ANTHROPIC_API_KEY as a repository secret in Settings → Secrets and variables → Actions.

---

Module Progression

Each module's exercise builds on the previous. By Module 8 you have a fully assembled production agent — every component built in Modules 1–7 is integrated into the capstone.

Module	Topic	Key Output
1	Hello Agent	First Claude API call, parse JSON
2	First AI Agent	Five-step agentic loop
3	ReAct & Planning	Multi-step iterative reasoning
4	Diagnosis & Remediation	Event-driven triage agent
5	Intelligent CI/CD	Quality gate with release decision
6	Conversational Observability	Natural-language ops queries
7	Multi-Agent Coordination	Orchestrator + specialist routing
8	Capstone	Full production agent pipeline