Workshops: Agentic Coding Tools in Depth
Claude Code Hooks, Skills & Plugins: Advanced Team Workshop
Most teams run Claude Code with its defaults, and every developer tunes it differently. In this advanced workshop your team turns Claude Code into a shared, versioned tool: hooks that enforce rules automatically, skills and slash commands that encode workflows, plugins that distribute them, and permissions your security team can sign off.
Who is this workshop for?
- Developers who already use Claude Code daily and want more than the defaults
- Tech leads responsible for team conventions and code quality
- Platform and developer experience engineers who maintain shared tooling
- Security champions who define what the agent may run
- Teams that want one consistent setup across many repositories
What you will take away
- Write hooks that run formatters, linters or tests after edits and block risky commands before they execute.
- Turn recurring workflows into custom slash commands and skills that Claude loads only when relevant.
- Package commands, skills, hooks, subagents and MCP configuration as a plugin for an internal marketplace.
- Structure user, project, local and managed settings so personal preferences never override company policy.
- Choose permission modes and allow and deny rules per repository and justify them to your security team.
Agenda
Module 1
How Claude Code is configured
- The settings hierarchy: user, project, local and managed policy
- What belongs in CLAUDE.md, in settings or in a skill
- Review of the participants' current setups
Module 2
Hooks as automatic guardrails
- Hook events: before and after tool use, prompt submission, session start, stop
- Exercise: auto-format and lint after every file edit
- Exercise: block destructive commands and edits to protected paths
Module 3
Slash commands and skills
- Custom slash commands with arguments for reviews or release notes
- Skills: instructions, scripts and reference files loaded on demand
- Exercise: a skill for your API or migration conventions
Module 4
Plugins and team distribution
- Bundling commands, skills, hooks, subagents and MCP servers into a plugin
- A private marketplace in a team repository
- Versioning, ownership and review of shared plugins
Module 5
Permissions and policy
- Permission modes from plan mode to auto-accepting edits, and why skipping all prompts belongs in sandboxes
- Allow and deny rules for tools, commands, files and MCP servers
- Managed settings developers cannot override
From personal setup to team standard
Claude Code reads its configuration from several layers, and that is exactly where teams lose consistency: one developer has a clever hook, another allows far more commands, and nobody is sure which CLAUDE.md instructions apply. We map what lives where and agree on a target structure for your repositories.
The result is a small, reviewed configuration package that new team members get automatically when cloning, instead of a wiki page of tips.
Hooks are deterministic, prompts are not
An instruction in CLAUDE.md is a request the model usually follows. A hook is code that runs every time its event fires. That decides where a rule belongs: style preferences can live in memory files, but anything that must never happen, such as editing migration files, touching credentials or pushing to main, needs a hook or a deny rule.
- Deterministic checks: formatting, linting, type checks after edits
- Hard blocks: dangerous commands, protected directories, credential files
- Context at session start: current ticket, branch or environment
Prerequisites and format
Participants should already work with Claude Code; this is not an introduction. neooutput AI Code Engineers lead the day in your repositories or a sandbox copy, remotely or on-site, in English or German. Your team leaves with a pull request containing its first shared configuration.
Frequently asked questions
Do we need prior Claude Code experience?
Yes. We assume participants use Claude Code regularly and know CLAUDE.md and basic permissions. Newcomers should start with our introductory Claude Code workshop, or we combine both into two days.
Are hooks and plugins a security risk themselves?
They can be: hooks run shell commands with the developer's rights, and plugins may bring their own hooks and MCP servers. We cover review rules for shared configuration and how to keep plugins from outside sources under control.
Can we reuse this setup with other coding agents?
Partly. Skills are plain folders with Markdown and scripts, MCP servers work with many clients, and good project instructions transfer easily. Hooks and permission settings are specific to Claude Code.
Workshop
Get in touch
Ready to turn your team's Claude Code tricks into a shared, enforceable setup?
Tell us roughly how many people will join and pick dates that work – we will get back to you within 24 hours with a proposal for content, format and schedule.
Prefer to talk first?
15 minutes, free and without obligation – with our AI Trainers and AI Code Engineers.
Book a call for free (opens in new tab) +49 761 519 751 50 (24/7)Request a training / workshop
Workshop: Claude Code Hooks, Skills & Plugins: Advanced Team Workshop · 1 day
More workshops
- Workshops: Agentic Coding Tools in DepthClaude Code Subagents & Git Worktrees: Parallel Agents Workshop
- Workshops: Agentic Coding Tools in DepthClaude Code in CI/CD: GitHub Actions & Headless Mode Workshop
- Workshops: Agentic Coding Tools in DepthClaude Agent SDK Workshop: Build Custom Engineering Agents
- Workshops: Agentic Coding Tools in DepthAdvanced OpenAI Codex Workshop: Cloud Tasks, CLI & Code Review
- Workshops: Agentic Coding Tools in DepthAdvanced Cursor Workshop: Project Rules, Background Agents & Bugbot
- Workshops: Prompt EngineeringPrompting Image AI: Midjourney, ChatGPT, Firefly & Gemini
