Claude Code Toolkit

A comprehensive collection of agents, skills, commands, rules, hooks, and utilities to supercharge your Claude Code experience.

What's inside

Component	Count	Description
Agents	31	Specialized agents for code review, TDD, security, architecture, and more
Skills	67	Domain knowledge packs with scripts, references, and templates
Commands	17	Slash commands for common workflows (/tdd, /plan, /build-fix...)
Rules	17	Coding standards for 12+ languages and cross-cutting concerns
Hooks	1	Telegram approval hook for remote action control
Statusline	1	Rich status bar with context usage, tokens, costs, and git info

Quick Install

git clone https://github.com/webcomunicasolutions/claude-code-toolkit.git
cd claude-code-toolkit
bash install.sh

Install individual components:

bash install.sh --agents      # 31 agents
bash install.sh --skills      # 67 skills
bash install.sh --commands    # 17 commands
bash install.sh --rules       # 17 coding rules
bash install.sh --statusline  # Status bar
bash install.sh --hook        # Telegram approval hook
bash install.sh --ecc         # Import extras from Everything Claude Code

Everything Claude Code (ECC) Integration

The installer also pulls additional components from the Everything Claude Code community repository. This adds 200+ extra skills, additional agents, commands, and hooks on top of what this toolkit provides.

• ECC components are only imported if they don't already exist locally (no overwrites)
• Run bash install.sh --ecc at any time to pull new ECC components
• Run bash install.sh --ecc-only to only sync ECC without reinstalling the toolkit
• Run bash install.sh --summary to see total installed component counts

Post-Install Configuration

Enable Agent Teams

Agent teams allow Claude Code to spawn specialized sub-agents for parallel work. Add to your ~/.claude/settings.json:

{
  "agentTeam": {
    "enabled": true
  }
}

With teams enabled, Claude can delegate tasks to the installed agents (code reviewers, build resolvers, TDD agents, etc.) and run them in parallel for faster results.

Verify Installation

After installing, restart Claude Code and run:

bash install.sh --summary

This shows the total count of agents, skills, commands, and rules installed.

Components

Agents (31)

Code Reviewers

Agent	Languages/Focus
code-reviewer	General quality, security, and maintainability
typescript-reviewer	TypeScript/JavaScript, React, Next.js
python-reviewer	Python, Django, FastAPI, Flask
go-reviewer	Go concurrency, error handling, idioms
rust-reviewer	Rust safety, ownership, idioms
java-reviewer	Java, Spring Boot, JPA
kotlin-reviewer	Kotlin, Android/KMP, Compose
cpp-reviewer	C++ memory safety, RAII, modern C++
flutter-reviewer	Flutter/Dart, widgets, state management
database-reviewer	PostgreSQL queries, schemas, RLS, indexing

Build Resolvers

Agent	Scope
build-error-resolver	Generic build/compilation errors
cpp-build-resolver	C++, CMake
go-build-resolver	Go, go vet, linters
java-build-resolver	Java, Maven, Gradle
kotlin-build-resolver	Kotlin, Gradle
rust-build-resolver	Rust, cargo, borrow checker
pytorch-build-resolver	PyTorch, CUDA, tensor shapes

TDD Cycle

Agent	Phase
tdd-red	Write failing tests first
tdd-green	Minimum code to pass tests
tdd-refactor	Improve quality keeping tests green
test-generator	Comprehensive test case generation

Architecture & Planning

Agent	Purpose
architect	System design and trade-off evaluation
principal-engineer	Engineering leadership, Martin Fowler style
planner	Break features into actionable steps
devils-advocate	Challenge assumptions, force critical thinking

Operations

Agent	Purpose
security-reviewer	Vulnerability analysis before deployments
verifier	Evidence-based verification gate
doc-updater	Keep documentation in sync with code
refactor-cleaner	Dead code, duplicates, unused dependencies
loop-operator	Autonomous work loops with safety guardrails
web-scraper-expert	Web scraping with anti-bot bypass

Commands (17)

Slash commands that activate specialized modes:

Command	Description
/optimize	Agent-driven project audit (diagnose, execute, verify)
/n8n	Activate n8n specialist mode
/tdd	TDD workflow: RED/GREEN/REFACTOR cycle
/plan	Create implementation plan before coding
/build-fix	Systematic build error resolution
/code-review	Exhaustive code review with security checklist
/devfleet	Orchestrate multiple agents in parallel with worktrees
/loop-start	Start autonomous loop with safety guardrails
/loop-status	Check active loop status
/checkpoint	Create code state checkpoint
/save-session	Save current session state
/resume-session	Resume a previously saved session
/model-route	Select optimal model per task and budget
/eval	Define and run feature evaluations
/harness-audit	Audit Claude Code harness configuration
/autoresearch	Autonomous improvement loop (Karpathy's AutoResearch pattern)
/full	Full autonomy mode - use all resources without asking

Skills (67)

Document Generation

Skill	Description
pdf	PDF manipulation: extract, create, merge, split, forms
docx	Word documents with tracked changes and comments
pptx	Presentations with layouts and speaker notes
xlsx	Spreadsheets with formulas, formatting, analysis
documentos-corporativos	Corporate PDF templates with header/footer branding

n8n Workflow Automation

Skill	Description
n8n-workflow	Create, modify, and optimize n8n workflows
n8n-workflow-patterns	Proven architectural patterns for workflows
n8n-code-javascript	JavaScript in n8n Code nodes
n8n-code-python	Python in n8n Code nodes
n8n-expression-syntax	Expression syntax and troubleshooting
n8n-mcp-tools-expert	MCP tools for n8n management
n8n-node-configuration	Node configuration guidance
n8n-validation-expert	Validation error interpretation

Language Patterns & Testing

Skill	Description
python-patterns	Idiomatic Python: type hints, dataclasses, async
python-testing	pytest fixtures, parametrization, mocking
golang-patterns	Go idioms: error handling, concurrency, interfaces
golang-testing	Table-driven tests, subtests, benchmarks
rust-patterns	Ownership, error handling, traits, async
rust-testing	Unit, integration, property-based testing
kotlin-testing	Kotest, MockK, coroutines, Flow

Web Frameworks

Skill	Description
django-patterns	Django/DRF: models, views, serializers, signals
laravel-patterns	Laravel: controllers, Eloquent, queues, events
springboot-patterns	Spring Boot: REST, JPA, validation, caching

Frontend & Design

Skill	Description
frontend-design	Production-grade web interfaces with high design quality
web-artifacts-builder	Multi-component HTML artifacts with React + Tailwind + shadcn/ui
canvas-design	Visual art in PNG/PDF using design principles
algorithmic-art	Generative art with p5.js and seeded randomness
brand-guidelines	Anthropic brand colors and typography
theme-factory	10 pre-set themes for any artifact
slack-gif-creator	Animated GIFs optimized for Slack

DevOps & Infrastructure

Skill	Description
docker-patterns	Docker/Compose for dev and production
postgres-patterns	PostgreSQL indexing, RLS, optimization
database-migrations	Safe, reversible, zero-downtime migrations
linux-server-audit	Complete server health check and security audit

AI & Automation

Skill	Description
agentic-engineering	Multi-agent system design patterns
autonomous-loops	Autonomous Claude Code execution patterns
cost-aware-llm	LLM API cost control and optimization
workflow-automation	n8n, Temporal, Inngest patterns
ollama-specialist	Local LLM management with GPU
prompt-optimizer	6-phase prompt improvement pipeline
whisper-transcribe	Audio/video transcription with faster-whisper
youtube-transcript	YouTube video transcript extraction

Testing & QA

Skill	Description
e2e-testing	Playwright E2E with Page Object Model
playwright-cli	Browser automation for testing and scraping
webapp-testing	Local web app testing with Playwright
test-and-fix	Autonomous QA loop: test all pages, fix bugs, repeat until clean
verification-loop	6-phase QA system before delivery
adversarial-review	Adversarial reviewer loop that fixes and retests until zero bugs
quality-gate	Combines all reviewers in a loop until zero bugs
dev-browser	Browser automation with persistent page state
screenshot-compare	Screenshot capture for visual comparison

Project Management

Skill	Description
project-optimizer	6-phase project audit and optimization
project-bridge	Structured communication between Claude Code projects
session-management	Session persistence between conversations
doc-maintenance	Documentation structure and organization
doc-coauthoring	Structured doc co-authoring workflow
doc-specialist	Generate agents from scraped documentation
skill-creator	Guide for creating new skills
skills-sh-search	Browse and install skills from skills.sh
spec	Spec-driven kickoff for large tasks (Karpathy 3-layer method)

Communication

Skill	Description
internal-comms	Internal communications templates
telegram	Telegram approval mode toggle
web-scraper	Website scraping to clean Markdown

Specialized

Skill	Description
api-design	REST API conventions and design
mcp-builder	Build MCP servers (Python/Node)
llmwhisperer	LLMWhisperer v2 OCR API for document text extraction
critico	Always-on critical, honest attitude (no default validation)

Rules (17)

Coding standards and best practices loaded automatically:

Rule	Scope
coding-style	General: immutability, file size, function size, error handling
security	OWASP checks, secret management, input validation
testing	80% coverage minimum, TDD workflow, anti-patterns
git-workflow	Commit format, PR standards, branch naming
patterns	Repository pattern, API response format, error handling
performance	Model selection, context window, optimization
typescript	TS/JS: type safety, Zod validation, React patterns
python	PEP 8, type annotations, pytest, immutability
go	gofmt, error wrapping, functional options, table-driven tests
rust	rustfmt, clippy, error handling, unsafe boundaries
java	google-java-format, records, Spring Boot patterns
kotlin	ktlint/Detekt, sealed classes, coroutines, Result
cpp	C++17+, RAII, smart pointers, sanitizers
csharp	.NET conventions, nullable types, async/await
php	PSR-12, strict_types, Laravel patterns
perl	v5.36, Moo, taint mode, Test2::V0
swift	SwiftFormat/Lint, value types, actors, Keychain

Hooks

Telegram Approval

Smart PreToolUse hook that classifies operations by risk level and lets you approve dangerous ones via Telegram.

• 3 modes: Terminal only, Terminal + Telegram (hybrid), Telegram blocking
• Smart filtering: Safe ops auto-approve, dangerous ops require approval
• Configurable sensitivity: all, smart (default), critical

Statusline

Custom Statusline

Rich status bar showing real-time metrics:

• Context window usage with progress bar
• Input/output token counts
• Session tracking (prompts, burn rate)
• Git branch info

Requirements & Dependencies

Quick Setup

# Install the toolkit
bash install.sh

# Install all system dependencies
bash install-deps.sh --all

# Check what's installed
bash install-deps.sh --check

System Packages (apt)

Package	Used by	Purpose
jq	Hooks, Statusline	JSON processing
curl	Hooks, Web scraper	HTTP requests
bc	Statusline	Cost calculations
git	Multiple	Version control
pandoc	docx, pptx	Document conversion
poppler-utils	pdf, docx, pptx	PDF tools (pdftotext, pdftoppm)
libreoffice	docx, pptx, xlsx	Office document processing
qpdf	pdf	PDF merge/split
ffmpeg	whisper-transcribe	Audio/video processing
xvfb	screenshot-compare	Virtual display for headless browsers
python3 / pip3	Multiple skills	Python runtime
nodejs / npm	Multiple skills	Node.js runtime

Python Packages (pip)

Package	Used by
reportlab	pdf, documentos-corporativos
pypdf, pdfplumber	pdf
pytesseract, pdf2image	pdf (OCR)
openpyxl, pandas	xlsx
defusedxml	docx, pptx, xlsx
markitdown[pptx]	pptx
beautifulsoup4, lxml, html5lib	web-scraper
nodriver	screenshot-compare
pillow, imageio, numpy	slack-gif-creator
youtube-transcript-api	youtube-transcript
requests	Multiple

Node Packages (npm)

Package	Used by
playwright	e2e-testing, webapp-testing, screenshot-compare
docx	docx
pptxgenjs	pptx
sharp	pptx, web-artifacts-builder
react, react-dom, react-icons	web-artifacts-builder, pptx

Optional

Package	Size	Used by	Install
faster-whisper	~500MB	whisper-transcribe	bash install-deps.sh --whisper
Chromium browser	~200MB	Playwright skills	npx playwright install chromium
Telegram bot	-	Telegram hook	See hook README

Install by Category

bash install-deps.sh --system   # System packages (apt)
bash install-deps.sh --python   # Python packages (pip)
bash install-deps.sh --node     # Node packages (npm)
bash install-deps.sh --whisper  # Audio/video transcription (optional)
bash install-deps.sh --check    # Verify what's installed

Contributing

PRs welcome! If you have a useful agent, hook, skill, or rule for Claude Code, open a PR.

License

MIT