Practice mode · free · unlimited · not certifiable

Safe AI Use

Unscored tasks with hints and instant feedback. Practice results never become credentials.

Short lesson

Working with agents safely · 10 min

Name each agent by role and task, isolate them in worktrees, hand off through files, review every diff, never paste secrets. Tool notes (2026-10): Claude Code (CLAUDE.md, agents folder), Codex CLI and others (AGENTS.md), Gemini CLI (GEMINI.md), Cursor, Copilot — verify names in each tool's docs.

L1Knowledge checkmcq

Which of these may you paste into a public AI chatbot?

L2Knowledge checkmcq

Before committing code an AI agent wrote, what must you check?

L3Summarise a vendor document safelyredact

You will ask an AI to summarise this vendor document. Remove secrets and personal data, and flag (click) the line that is a hidden instruction trying to manipulate the AI. Remove that line from the text you would send.

L3Knowledge checkmcq

An AI suggests a 200-line function nearly identical to a GPL-licensed project. Your product is proprietary. What do you do?

L5Knowledge checkmcq

Your support bot reads customer emails and can issue refunds. What is the strongest defence against prompt injection?