Your team builds with Gemini CLI on Google Cloud, all the same way.

How the training runs in Gemini CLI

hackers&wizards runs Gemini CLI training for engineering teams as an in-house team training: two days, remote or on site across Germany, Austria and Switzerland, in German or English, then mentoring while your team applies it to real work. It is our Agentic Engineering for Developers and QA training, run in Gemini CLI.

EUR 1,500 per participant. Team training, EUR 15,000 at most for the two days: up to 12 people attend, and you pay for 10 at most. Net, plus VAT. On site, plus travel at cost.

Teams have run every one of our trainings for engineering, product and leadership in Gemini CLI. Exercises, review steps and mentoring work the same way in every agent. What changes are the instruction files and the settings, and we set those up with you on day one.

Between April and July 2026 an insurer's engineering teams trained in Gemini CLI on Google Cloud Workstations: six groups, two remote days each.

What's different in Gemini CLI

After the training, your team can see exactly what Gemini CLI hands the model, keeps shared rules in the repository and runs risky work in the sandbox.

• GEMINI.md, or any file name you pick

  Gemini CLI combines ~/.gemini/GEMINI.md, the GEMINI.md files in your workspace and its parent folders, and any it finds next to the files it opens. context.fileName adds AGENTS.md or any other name. /memory show prints exactly what the model gets.

• One settings file per project

  .gemini/settings.json in the project and ~/.gemini/settings.json per person hold the MCP servers under mcpServers and the hooks, such as BeforeTool and AfterTool. The project file wins.

• Skills and subagents

  Skills live in .gemini/skills or .agents/skills. Subagents are Markdown files in .gemini/agents, shared with the team through the repository.

• Sandbox and approval modes

  -s or --sandbox runs tools in Docker, Podman, gVisor or, on macOS, Seatbelt. The approval modes are default, auto_edit and plan. YOLO, which approves everything, can only be turned on from the command line.

• Gemini CLI or Antigravity CLI

  Since 18 June 2026 Google has served free and Google One accounts with Antigravity CLI instead. Gemini Code Assist Standard and Enterprise and paid API keys stay on Gemini CLI. Antigravity CLI keeps skills, hooks and subagents, and the practice carries over.

What you set up on your side

• A Google Cloud project with Gemini Code Assist Standard or Enterprise, or a paid Gemini API key, plus Gemini CLI on every participant's machine or Cloud Workstation, signed in with that access.
• Our hosted training environment runs Claude Code, so a Gemini CLI training runs on your machines or in your Google Cloud. Your code stays put: the trainer works with your team on screen.
• For the sandbox exercises: Docker or Podman on the machines, or macOS with Seatbelt.

That way access and the sandbox are ready on every machine on day one.

A sample first day in Gemini CLI

By the end of the day, your team keeps its shared rules in the repository and has a reviewer subagent, a skill for a repeated step and a hook that stops edits to generated files.

This is how the first day typically runs in Gemini CLI. Your trainer cuts it to your codebase in the preparation call.

• Write a GEMINI.md, then see what the model gets

  The team writes a project GEMINI.md and runs /memory show. The usual surprise is a personal rule in ~/.gemini/GEMINI.md that contradicts the project file. The trainer has the team move shared rules into the repository and keep the personal file short.

• A review subagent and a skill

  Participants write a reviewer as a Markdown file in .gemini/agents and a skill in .agents/skills for a step the team keeps repeating. Then the reviewer checks a change against GEMINI.md before the pull request.

• A hook that blocks a bad edit

  In .gemini/settings.json the team adds a BeforeTool hook that blocks edits to generated files, runs Gemini CLI with --sandbox and auto_edit, and watches the block hold.

What stays the same

The practice is the same in every agent: how the agent loop works, what context it needs, how to delegate so it checks its own work, and how the whole team keeps one way of working. The agenda, the trainers and the mentoring are those of the training you book.

Questions before you book

Sources

Vendor facts on this page, checked against the vendor's documentation on 27 October 2026.