Get your tools ready. Meet your lab companion. Watch it build something live.
This week: deploy models, integrate APIs, build workflows, red team AI apps. Most of that is code.
What you'll do this week
The analogy: You wouldn't hand-craft iptables rules when you have Panorama. Same principle. You're the security architect. Claude Code is the implementer.
Not all of you write code daily. That's fine. Claude Code writes the code — you drive the intent. It's not a nice-to-have. It is the lab environment.
Agentic coding assistant in your terminal. Reads code, writes code, runs commands, iterates on failures.
Key differentiator: It's agentic. It doesn't suggest code — it executes, tests, sees failures, fixes, and retries. Autonomously.
Switching to terminal. Watch — no action required yet.
Takeaway: This is what you'll be able to do by the end of the week. Claude Code bridges the gap between your security expertise and the coding these labs require.
We need everyone in the lab GitHub org. Let's build the tool that does it.
Quick pulse check. Be honest — that's what the next 20 minutes are for.
Missing 3+? Flag me or a TA right now. We'll get you sorted. Don't wait.
Two sections. Work top to bottom. Don't spend more than 2-3 minutes stuck on anything.
Verify commands for every pre-work item. Don't assume — run the commands and confirm output.
git/gh CLI, GitHub org, vertex-search MCP, Python/Docker for model security lab.
Fix order if behind
Open the lab guide. Work through it. Raise your hand if stuck. Help your neighbor if done.
When done: Claude Code is your lab partner for the rest of the week.
Advanced session (Module 3.6) on Day 3 — today is just getting you set up.