tutorial_agent

Tutorial Agent (Go)

Go-based AI tutorial assistant for:

• building learning paths
• creating step-by-step lessons
• generating hands-on exercises
• saving each generated tutorial as a .md file under minikube/tutorial-ai/

1) Configure environment

cd agents/tutorial_agent
cp .env.example .env

Required:

• GITHUB_MODELS_TOKEN

Optional:

• GITHUB_MODELS_MODEL (default: openai/gpt-5-mini)
• GITHUB_MODELS_ENDPOINT (default: https://models.github.ai/inference/chat/completions)
• PORT (default: 8088)
• TUTORIAL_OUTPUT_DIR (optional override for markdown output folder)

2) Run

From repo root:

go run ./agents/tutorial_agent

3) API

Health

curl http://localhost:8088/health

Tutorial

curl -X POST http://localhost:8088/tutorial \
  -H "Content-Type: application/json" \
  -d '{
    "mode": "lesson",
    "topic": "kubernetes deployments",
    "level": "beginner",
    "outputFile": "k8s-deployments-lesson.md",
    "message": "Teach me with an example and a mini practice task"
  }'

The agent writes a markdown file in minikube/tutorial-ai/ by default and returns:

• filePath
• fileName
• savedIn
• timestamp

mode options:

• plan
• lesson
• exercise
• default: general tutorial guidance

4) VS Code task

Run task: Run Tutorial Agent HTTP Server from agents/tutorial_agent/.vscode/tasks.json.