Skip to content

README.md

Latest commit

 

History

History
75 lines (55 loc) · 3.3 KB

README.md

File metadata and controls

75 lines (55 loc) · 3.3 KB
Raw

The labs

Two courses teach the same five labs on two stacks. Pick one.

Cloud course Local course
Runs on StreamNative Cloud: your own instance, with a Kafka cluster, a SQL workspace, and an agent workspace Your laptop: Ursa for Kafka, RisingWave, and the Orca Agent Engine
You need A StreamNative Cloud login with your own instance, from the hackathon organizers Docker and an Anthropic API key
Time About 40 minutes About 45 minutes, plus the image downloads
Take it when You are at the event You have no StreamNative Cloud instance, or you want to see every part run

Both courses use the same three paths for the agent steps. Pick one path and stay on it.

The labs

Lab You The idea
0. Set up Get your stack ready and run the doctor Know that every part answers before you build on it
1. Hello, agent Create an agent and talk to it Agent, environment, session, events
2. Hello, streaming SQL Build a materialized view over the login topic Context that keeps itself fresh
3. Agent + live context Give the agent SQL tools, then inject new data The answer changes with the data
4. Agent acts, human approves Let the agent write, with your OK Governed actions

The labs build on each other, so take them in order.

How a lab works

Every lab has the same parts, in the same order.

  1. Before you start says what has to be true first.
  2. Three or four steps. Each step has you do one thing, then explains what happened. Each step ends in a Check.
  3. Check your understanding: two or three questions. The answer is folded under each one; decide first, then open it.
  4. Try it yourself: a small task with no instructions, its own check, and a folded solution.
  5. Recap and What's next.

In a guided session the steps are what you do together. The questions and the task are yours to do afterwards.

How to check your work

A check is a command that reads something and changes nothing. Before you do a step, its check prints nothing, or says what is missing. After the step, it prints the evidence.

Most checks look like this:

./lab-ork agent get @agent_id -o json | jq -e '{name, version}'
  • ./lab-ork is ork, pointed at your Agent Engine with the key from your .env. It replaces @agent_id, @environment_id, @vault_id and @session_id with the ids your lab scripts saved in .orca-state/, so you never copy an id by hand.
  • jq -e prints what the filter selects, and exits non-zero when it selects nothing. A check passes when it prints something.

The other checks are SQL: a SELECT that returns a row once the step is done.

Keep two terminals open. Run the steps in your path's folder (cli/, python/ or typescript/), and run the checks from the repository root.

Learn with a tutor

If you use a coding agent such as Claude Code, it can walk you through a course one step at a time, check your work with you, and quiz you. See Learn with the tutor.