n8n-claude-skills

mcp
Guvenlik Denetimi
Gecti
Health Gecti
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 17 GitHub stars
Code Gecti
  • Code scan — Scanned 3 files during light audit, no dangerous patterns found
Permissions Gecti
  • Permissions — No dangerous permissions requested
Purpose
This MCP server provides a collection of production-grade skills and patterns for Claude Code, designed to help developers build, debug, and ship n8n automation workflows accurately without hallucinating node types or syntax.

Security Assessment
Overall risk: Low. The tool is essentially a knowledge base of text patterns rather than an active script or service. The installation process uses a standard shell script to clone and copy these skill files locally. The automated code scan checked three files and found no dangerous patterns, hardcoded secrets, or requests for excessive permissions. Because it acts primarily as a prompt guide for an AI agent rather than a standalone application, it does not inherently execute arbitrary shell commands, make external network requests, or access sensitive data on its own.

Quality Assessment
Overall quality: Good. The project is actively maintained, with its most recent push happening today. It is properly licensed under the standard MIT license. Community trust is modest but present, indicated by 17 GitHub stars and the author's stated credentials as an n8n Verified Creator. The repository includes a comprehensive, well-documented README that clearly explains exactly what users are installing and how the skills interact with their workflows.

Verdict
Safe to use. It is a straightforward, low-risk prompt library that effectively extends AI capabilities without exposing your system to security vulnerabilities.
SUMMARY

Production Claude Code skills for n8n from a Verified Creator's 100+ workflows

README.md

n8n-claude-skills

Production-grade Claude Skills for building, debugging, and shipping n8n workflows — distilled from 100+ production workflows by an n8n Verified Creator.

Give Claude Code the instincts of a senior n8n engineer. No more hallucinated node types, broken expressions, or toy demos.

# 60-second install
npx n8n-claude-skills install

Then in Claude Code:

"Build me an n8n workflow that ingests leads from a webhook, enriches them via SerpAPI → Apollo waterfall, scores spam with a composite check, and writes qualified leads to Google Sheets."

Claude now builds a real, production-shaped workflow — correct node types, proper expressions, checkpoint tables, error branches, and retry logic. Not a Figma mockup of a workflow. A workflow.


What's in the box

Skill What it does When it fires
workflow-architect Designs n8n workflows from requirements. Picks the right nodes, structures branches, and wires error paths. "build a workflow that…", "design an automation for…"
chain-llm-pattern The Groq/OpenAI chain_llm pattern for multi-step LLM reasoning (entity extraction → categorization → scoring). "extract structured data with LLMs", "multi-step prompt chain"
enrichment-waterfall Multi-vendor enrichment waterfalls (SerpAPI → Hunter → Apollo → LLM fallback) with cost-aware fallbacks. "lead enrichment", "data waterfall", "Clay alternative"
mysql-checkpointing Production MySQL patterns: batch processing, duplicate prevention, resumable pipelines, dynamic table creation. "idempotent pipeline", "checkpoint table", "batch processing"
debug-workflow Systematic n8n debugging: expression syntax, pinned data, sub-workflow isolation, execution log reading. "why is my workflow failing", "debug this n8n error"

More shipping weekly. PRs welcome.


Why this exists

LLMs confidently hallucinate n8n. They invent node types (Googlesheets instead of Google Sheets), mix up expression syntax ({{$json["x"]}} vs ={{ $json.x }}), forget about sub-workflows, and ship demos that collapse the moment you plug in real data.

This pack teaches Claude the patterns that actually survive production — the ones we use in:

  • Real estate financial reporting (Google Sheets → Claude API → PDF → Gmail)
  • Multi-language call transcript pipelines (EN/ES/PT, Groq chain_llm)
  • Composite spam scoring with Twilio retry logic
  • 47-node batch processing with duplicate prevention
  • WhatsApp commerce agents with Stripe + Meta Cloud API

You get the instincts, not the code.


Install

Claude Code (recommended)

# Clone and install to your user skills directory
git clone https://github.com/masteranime/n8n-claude-skills.git
cd n8n-claude-skills
./install.sh

Claude Code picks up the skills automatically.

Manual (any Skills-compatible agent)

Copy any folder inside skills/ into your agent's skills directory. Each folder is a self-contained skill with a SKILL.md and optional reference files.

Works with: Claude Code, Cursor, Codex, Gemini CLI, Opencode, and anything following the Agent Skills spec.


Quick demo

Prompt:

Build an n8n workflow that receives Stripe checkout.session.completed webhooks, looks up the customer in our MySQL orders table, sends a WhatsApp confirmation via Meta Cloud API, and retries via Twilio SMS if WhatsApp fails.

Without this pack: Claude writes pseudocode, invents Whatsapp Node, forgets the retry branch, skips the idempotency check.

With this pack: Claude produces a 14-node workflow JSON you can paste into n8n — with webhook signature verification, a MySQL processed_webhooks checkpoint, proper Meta Cloud API messages endpoint, an IF branch on WhatsApp failure feeding into Twilio with exponential backoff, and an error trigger sub-workflow.

Full example: examples/stripe-whatsapp-fallback.json


Who built this

Muhammad Shaheer — n8n Verified Creator (100+ workflows), Claude Code daily user, 4+ years shipping production AI automation.

These skills are extracted from real client work — real estate finance pipelines, multi-language NLP, voice lead automation, lead enrichment waterfalls. If a pattern is in here, it shipped to a paying customer first.


Contributing

Have a battle-tested n8n pattern? Open a PR. See CONTRIBUTING.md for the skill format and review criteria.

Good additions:

  • Domain skills (e-commerce flows, real estate, healthcare)
  • Integration skills (specific SaaS tools)
  • Debug recipes for common footguns

License

MIT. Use it, fork it, ship it. Credit appreciated but not required.


⭐ Star the repo if this saved you an hour of debugging n8n expressions.

Yorumlar (0)

Sonuc bulunamadi