The handbook/Memory & Advanced Tools
SettingsMy workspaceOfficial docs ↗

Extend your workspace

Teach Claude
how you work.

Start with project instructions. Add reusable workflows and integrations only when they solve a real need.

STEP-BY-STEP HANDBOOK15 min introduction

Looking for the app menu? Start with Customize & Plugins. For a first reusable recipe, try Skills. This page keeps the deeper memory, hook, and agent reference together.

Which extension does what?

FeatureWhat it isExample
CLAUDE.mdProject guidance Claude reads.Build commands and architecture conventions.
Auto memoryInformation Claude retains between sessions.Useful project facts and previous troubleshooting.
SkillA reusable task recipe.A release-readiness checklist.
HookA program triggered by an event.Run a formatter after certain edits.
MCP serverA connection that exposes external tools or data.Read issues from an authorized tracker.
PluginA packaged collection of extensions.A curated set of development workflows.
SubagentA separate assistant context for a bounded task.Inspect test coverage and return findings.

Create a useful CLAUDE.md

Inside your project, run /init, then inspect the generated file. Keep facts concise and accurate. Replace guessed build commands with commands that actually work in the repository.

Example CLAUDE.md file
# Project guidance

## Purpose
A simple personal website for learning.

## Working rules
- Explain a plan before large changes.
- Make the smallest change that satisfies the request.
- Match the existing project style.
- Do not add unrelated features or dependencies.
- Never include secrets in code or logs.

## Verification
- Review your own diff after every implementation or code change.
- Run the relevant checks supported by this project.
- For visible changes, inspect the page and keyboard behavior.
- Report checks performed, results, and anything untested.

## Delivery
- Summarize changed files and the reason for each change.
- Ask before publishing, merging, or changing external accounts.

Use /memory to inspect and manage instruction files and auto memory. Repository instructions can travel with Git; personal machine settings and memories do not automatically travel to cloud sessions.

Instructions guide behavior; permissions enforce access

A sentence in CLAUDE.md is not a sandbox or a guaranteed security boundary. Apply tool permissions, account restrictions, and operating-system isolation separately.

Create your first custom command

The Skills practice lesson contains the review-only recipe. Keep one copy there rather than maintaining overlapping examples. A project skill belongs in .claude/skills/; older .claude/commands/*.md files remain supported.

Connect tools with MCP

MCP means Model Context Protocol. A server offers tools Claude can call, such as querying an issue tracker. Some servers run locally; others are remote services. Neither the protocol nor a marketplace listing proves a server is trustworthy.

Choose one trusted integration

Read its official setup instructions and permissions. Identify whether it can only read or also change data.

Add it using its documented setup

Use the service’s exact endpoint or package. Inspect claude mcp --help for CLI options. Avoid copying an unknown server command from a forum.

Authenticate and inspect

Open /mcp, sign in if needed, and check the available tools. Start with the smallest useful account scope.

Test a harmless read

Ask it to list one non-sensitive item. Verify the response against the original service before granting write workflows.

Hooks and plugins

Hooks react to lifecycle events such as PreToolUse or PostToolUse. They can run real programs, so inspect the command and its inputs. Start with a hook that reports information before allowing one to modify files or block actions. Use /hooks to review configuration.

Use /plugin to inspect available plugin management. Prefer known publishers and review the package’s skills, hooks, and MCP connections. Removing a plugin is separate from revoking a credential it used.

Subagents, worktrees, and teams

Ask Claude to create a narrowly scoped reviewer or researcher when separate context is useful. Agent definitions may live in .claude/agents/ for a project or ~/.claude/agents/ for your machine. Separate worktrees provide separate working copies of a Git repository; they are not complete security isolation.

Agent teams coordinate multiple sessions and are an advanced, availability-dependent feature. More agents can increase cost and conflicting edits. Start with one assistant, then use parallel work only for tasks with clear boundaries.

Was this chapter useful?

Report an error

Votes are shared only if you enable optional measurement in Privacy & Preferences.

CONTINUE THE GUIDEAutomation & agents

Edit this example

Changes here affect this copy only. Nothing runs from this site. Closing this window discards your edits.

Search the field guide

Search all chapters and FAQ answers. Use Tab to reach a result and Enter to open it.