Your team hands Claude Code whole features, and the code holds up in review.

How the training runs in Claude Code

hackers&wizards runs Claude Code 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 Claude Code.

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.

We have worked with Claude Code since months before its public preview. Teams have run every one of our trainings for engineering, product and leadership in Claude Code.

In January 2026 a software company's development team trained two days in Claude Code on its own code: day one agentic development, day two feedback loops, subagents and the whole development lifecycle. Mentoring followed, half a day a week from April to June.

What's different in Claude Code

After the training, your team knows which instruction belongs in CLAUDE.md, a path rule or a skill, and sets permissions so routine commands run and sensitive ones are blocked.

• CLAUDE.md, loaded folder by folder

  When it starts, Claude Code loads the CLAUDE.md in its working folder and in every folder above it. A subfolder's CLAUDE.md joins as soon as Claude reads or edits a file in that folder. The docs recommend staying under 200 lines per file. @path imports split a long file but don't make it any cheaper. If a repository has only an AGENTS.md, Claude Code reads that instead.

• Rules scoped to paths

  .claude/rules holds one Markdown file per topic. Give a rule a paths list in its frontmatter and it loads only when Claude works on matching files. Your testing rules stay out of a CSS change, and CLAUDE.md stays short.

• Skills and subagents in .claude

  A skill is a folder in .claude/skills with a SKILL.md inside. Claude sees every skill's description and loads the full skill only when it's used or someone types /skill-name. A subagent is a Markdown file in .claude/agents with its own prompt, tools and model, and it runs in its own context window.

• Hooks and MCP live in the repo

  Hooks go under hooks in .claude/settings.json and fire on events like PreToolUse, PostToolUse and Stop. A PreToolUse hook that exits with code 2 blocks the call. MCP servers the whole team uses go into .mcp.json via claude mcp add --scope project, and Claude Code asks before using one for the first time.

• Permission modes and permission rules

  Shift+Tab cycles the modes. Manual asks before edits and commands, acceptEdits lets file edits through, plan reads and plans before anything changes, and auto runs everything with safety checks in the background. Rules in .claude/settings.json are evaluated deny, then ask, then allow, so a deny rule beats any allow rule.

What you set up on your side

• Claude Code on every participant's machine, in the terminal, VS Code or a JetBrains IDE, signed in with your Claude plan or your company's API access. Or use our hosted training environment, a browser editor with Claude Code already in it, so there's nothing to install before day one.
• Auto mode is available by default on Team and Enterprise plans, and an admin can turn it off in managed settings. Before day one we check with you which permission modes your policy allows.
• In the prep call we decide whether your team practises on one of our sample repositories or its own. The day before, we check every participant's setup.

That way every participant is working in Claude Code before lunch on day one.

A sample first day in Claude Code

By the end of the day, your team plans a feature with Claude before any file changes, and leaves with a CLAUDE.md, a skill and a read-only reviewer of its own.

Your trainer cuts the day to your codebase in the preparation call.

• From /init to a CLAUDE.md worth keeping

  Participants start Claude Code in the repository, ask what the project does and use /context to see what's filling the context window. /init drafts a CLAUDE.md, and the team trims it to what Claude can't work out from the code. After /clear, asking the same question again shows what the file changed.

• Plan mode first, then build

  Shift+Tab switches to plan mode: Claude reads the codebase and writes a plan before touching a file. The team reviews the plan, builds the feature with edits accepted, and later picks the session back up with --resume.

• A skill and a read-only subagent

  In the afternoon the team packages the morning's steps as a skill with phases: research, plan, test, refactor, commit. Alongside it goes a refactoring advisor in .claude/agents, with read-only tools and its own model, that reviews the result before the pull request.

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 29 October 2026.