Claude Code Cookbook

April 14, 2026 · View on GitHub

A collection of commands, roles, and automation scripts for Claude Code.

Automate your workflow without unnecessary confirmations, allowing you to focus on what matters. Claude Code judges and executes common tasks like code fixes, test runs, and documentation updates.

Read in your language: 🇯🇵 日本語 · 🇺🇸 English · 🇰🇷 한국어 · 🇨🇳 简体中文 · 🇹🇼 繁體中文 · 🇪🇸 Español · 🇫🇷 Français · 🇧🇷 Português

What is Claude Code Cookbook?

Claude Code Cookbook provides a plugin system that extends Claude Code with:

  • Commands: Custom slash commands for common development tasks
  • Roles: Expert role presets for specialized assistance
  • Hooks: Automated scripts that trigger at specific events

Key Features

Commands

39 slash commands organized by category. Execute by typing / followed by the command name.

Pull Request Management

CommandDescription
/pr-listDisplay prioritized list of open PRs in current repository
/pr-issueDisplay prioritized list of open Issues in current repository
/pr-createAuto-generate PR from Git changes with detailed description and optimal branch setup
/pr-reviewSystematic code quality and architecture review for Pull Requests
/pr-fixEfficiently respond to PR review comments with 3-stage error analysis
/pr-auto-updateAutomatically update PR description and labels based on changes

Code Quality & Refactoring

CommandDescription
/refactorSafe, step-by-step code refactoring with SOLID principles evaluation
/smart-reviewAdvanced code review to improve quality
/tech-debtAnalyze technical debt and create prioritized improvement plans
/analyze-dependenciesAnalyze project dependencies and visualize circular dependencies
/analyze-performanceAnalyze application performance issues and propose improvements
/design-patternsPropose and review implementations based on design patterns

Development Tools

CommandDescription
/fix-errorSuggest code fixes based on error messages
/explain-codeClearly explain functionality and logic of selected code
/commit-messageGenerate commit messages based on changes
/semantic-commitSplit large changes into meaningful units with semantic commit messages
/pr-checksMonitor GitHub Actions CI status and track until completion
/screenshotCapture and analyze screen screenshots

Planning & Analysis

CommandDescription
/planActivate planning mode and formulate detailed implementation strategies
/specCreate detailed specifications from requirements (spec-driven development)
/ultrathinkExecute structured thinking process for complex issues
/check-factVerify information accuracy by referencing codebase and documentation
/sequential-thinkingThink through complex problems step-by-step using Sequential Thinking MCP
/taskLaunch specialized agents for autonomous search, research, and analysis tasks

Dependency Management

CommandDescription
/update-node-depsSafely update dependencies in Node.js projects
/update-flutter-depsSafely update dependencies in Flutter projects
/update-rust-depsSafely update dependencies in Rust projects

See your language plugin for the complete list of 39 commands with detailed documentation.

Roles

Switch Claude to expert roles for specialized assistance. Each role can run independently as a sub-agent using the --agent option for parallel execution without interfering with main context.

RoleDescription
securitySecurity vulnerability analysis and threat detection
architectSoftware architecture design and system design patterns
frontendUI/UX optimization and frontend best practices
backendAPI design, microservices, and cloud-native architecture
performancePerformance optimization for speed and memory
qaTest planning and quality assurance strategies
mobileiOS/Android development and mobile-first design
reviewerCode review focusing on readability and maintainability
analyzerSystem analysis and root cause analysis

Usage Examples

# Normal mode (execute in main context)
/cook:role security
"Review this authentication system for vulnerabilities"

# Sub-agent mode (execute in independent context)
/cook:role security --agent
"Perform comprehensive security audit of entire project"

# Multiple roles in parallel
/cook:multi-role security,performance --agent
"Analyze system's security and performance comprehensively"

Hooks

Automate development workflow with event-triggered scripts configured in settings.json:

Hook ScriptEventDescription
preserve-file-permissions.shPreToolUse / PostToolUseSave and restore file permissions to prevent unintended changes

Development Workflow

Typical development flow using Claude Code Cookbook commands:

flowchart TB
    Start([Task Confirmation]) --> PRList["/cook:pr-list<br/>Open PR List"]
    Start --> PRIssue["/cook:pr-issue<br/>Open Issue List"]

    PRList --> TaskType{Type?}
    PRIssue --> TaskType

    TaskType -->|New Feature| Plan["/cook:spec<br/>Requirements & Design"]
    TaskType -->|Bug Fix| Fix["/cook:fix-error<br/>Error Analysis"]
    TaskType -->|Refactoring| Refactor["/cook:refactor<br/>Improvement"]
    TaskType -->|Review| Review["/cook:pr-review<br/>Review"]

    Plan --> Design["/cook:role architect<br/>/cook:role-debate<br/>Design Consultation"]
    Design --> Implementation[Implementation & Testing]
    Fix --> Implementation
    Refactor --> Implementation
    Review --> Implementation

    Implementation --> Check["/cook:smart-review<br/>Quality Check"]
    Check --> Commit["/cook:semantic-commit<br/>Commit by Purpose"]
    Commit --> PR["/cook:pr-create<br/>Auto PR Creation"]
    PR --> CI["/cook:pr-checks<br/>CI Status Check"]

    CI --> Status{Issues?}
    Status -->|Yes| Feedback["Fix Response<br/>/cook:pr-fix<br/>/cook:fix-error"]
    Status -->|No| End([Complete])

    Feedback --> Implementation

    classDef commandBox fill:#e0f2fe,stroke:#0369a1,stroke-width:2px,rx:5,ry:5,color:#0c4a6e
    classDef processBox fill:#f0f9ff,stroke:#0ea5e9,stroke-width:1px,rx:5,ry:5,color:#075985
    classDef decisionBox fill:#fef3c7,stroke:#f59e0b,stroke-width:2px,rx:5,ry:5,color:#78350f
    classDef startEnd fill:#86efac,stroke:#22c55e,stroke-width:2px,rx:20,ry:20,color:#14532d

    class PRList,PRIssue,Plan,Fix,Refactor,Review,Design,Check,CI,Commit,PR,Feedback commandBox
    class Implementation processBox
    class TaskType,Status decisionBox
    class Start,End startEnd

Installation

Step 1: Add the Marketplace

First, add this repository as a plugin marketplace in Claude Code:

/plugin marketplace add wasabeef/claude-code-cookbook

Step 2: Install Your Language Plugin

Choose and install your preferred language plugin:

LanguagePlugin NameInstall Command
🇯🇵 日本語plugins/ja/plugin install cook@claude-code-cookbook
🇺🇸 Englishplugins/en/plugin install cook-en@claude-code-cookbook
🇰🇷 한국어plugins/ko/plugin install cook-ko@claude-code-cookbook
🇨🇳 简体中文plugins/zh-cn/plugin install cook-zh-cn@claude-code-cookbook
🇹🇼 繁體中文plugins/zh-tw/plugin install cook-zh-tw@claude-code-cookbook
🇪🇸 Españolplugins/es/plugin install cook-es@claude-code-cookbook
🇫🇷 Françaisplugins/fr/plugin install cook-fr@claude-code-cookbook
🇧🇷 Portuguêsplugins/pt/plugin install cook-pt@claude-code-cookbook

Each plugin includes:

  • Full command documentation in the native language
  • Role definitions optimized for that language
  • Language-specific customization examples

Quick Start

After installation, commands and agents are invoked with a plugin-specific prefix:

# Commands (example with Japanese plugin)
/cook:pr-create
/cook:semantic-commit
/cook:check-fact

# Agents
@agent-cook:frontend
/cook:role security

# Explore available commands
/cook:

Note: The plugin name is automatically added as a prefix to prevent namespace conflicts when multiple language plugins are installed.

Why Claude Code Cookbook?

  • Multilingual: 8 language plugins with native translations
  • Production-ready: Battle-tested commands and roles
  • Comprehensive: 39 commands covering entire development workflow
  • Extensible: Easy to add your own commands and roles
  • Safe: Built-in safety hooks prevent dangerous operations
  • Efficient: Reduce repetitive tasks and context switching
  • Open Source: Apache 2.0 licensed

Example Workflow

# 1. Check current tasks
/cook:pr-list

# 2. Plan new feature
/cook:spec
/cook:role architect

# 3. Implement with quality checks
# ... write code ...
/cook:smart-review

# 4. Commit changes logically
/cook:semantic-commit

# 5. Create PR automatically
/cook:pr-create

# 6. Monitor CI and respond to feedback
/cook:pr-checks
/cook:pr-fix