n8n-skills
A comprehensive skillset for Claude Code to build production-ready n8n workflows efficiently.

Suggested prompts
About this skill
n8n-skills
A set of 14 complementary Claude Code skills (plus a router skill and hooks enforcement layer) that teach AI assistants how to build production-ready n8n workflows using the n8n-mcp MCP server, and how to deploy self-hosted n8n.
Suitable Scenarios
- Developers building n8n workflows with the n8n-mcp MCP server.
- Users who need to learn correct n8n expression syntax, effective MCP tool usage, proven workflow patterns, and proper node configuration.
- Those deploying self-hosted n8n on a Linux VM.
Core Capabilities
The pack includes 14 skills that activate automatically based on your query. Each skill addresses a specific aspect of n8n workflow building:
- n8n Expression Syntax – Teaches correct n8n expression syntax and common patterns, including core variables ($json, $node, $now, $env) and common mistakes catalog.
- n8n MCP Tools Expert (Highest Priority) – Expert guide for using n8n-mcp MCP tools effectively, including tool selection, nodeType formats, validation profiles, smart parameters, and auto-sanitization.
- n8n Workflow Patterns – Build workflows using 5 proven architectural patterns: webhook processing, HTTP API, database, AI, scheduled. Includes a pattern selection guide.
- n8n Validation Expert – Interprets validation errors and guides fixing, with a real error catalog, false positives guide, and profile selection.
- n8n Node Configuration – Operation-aware node configuration guidance, including property dependency rules, operation-specific requirements, and AI connection types.
- n8n Code JavaScript – Write effective JavaScript code in n8n Code nodes, covering data access patterns, critical gotchas (webhook data under $json.body), correct return format, and top 5 error patterns.
- n8n Code Python – Write Python code in n8n Code nodes with proper limitations awareness, including data access, standard library reference, and workarounds for missing libraries.
- n8n Code Tool – Write code for the AI‑agent‑callable Custom Code Tool, which is a different node from the regular Code node with a different contract (return format is a string, not [{json:{...}}]).
- n8n Error Handling – Make failures loud, structured, and recoverable: per-node error outputs, retryOnFail, 4xx/5xx response shaping, and Error Trigger workflows.
- n8n Binary & Data – Handle files, images, and binary data correctly: $binary vs $json, keeping binary alive across JSON transforms, agent-tool binary boundary, and CDN/URL requirement for chat surfaces.
- n8n Sub-workflows – Build reusable, composable sub-workflows: Execute Workflow Trigger with typed inputs, mode: all vs each, waitForSubWorkflow, verb-first naming, and stateless vs stateful design.
- n8n AI Agents – Design n8n AI agents correctly: agent vs LLM Chain vs Text Classifier, tool names & descriptions as prompts, structured output, memory, human-in-the-loop, and chat topology.
- n8n Multi-Instance – Target the right n8n instance when an account has more than one: list/switch shapes, per-session binding, verifying current before credential writes, and recovering from misroutes.
- n8n Self-Hosting (deployment/ops) – Deploy production self-hosted n8n end-to-end to a fresh Linux VM using Docker Compose behind Caddy with automatic HTTPS, single or queue mode, secure defaults, and Day‑2 operations.
How It Works
The skills are designed to activate automatically when you ask a relevant question. For example:
- "How do I write n8n expressions?" → Activates: n8n Expression Syntax
- "Find me a Slack node" → Activates: n8n MCP Tools Expert
- "Build a webhook workflow" → Activates: n8n Workflow Patterns
- "Why is validation failing?" → Activates: n8n Validation Expert
- "How do I configure the HTTP Request node?" → Activates: n8n Node Configuration
- "How do I access webhook data in a Code node?" → Activates: n8n Code JavaScript
- "Can I use pandas in Python Code node?" → Activates: n8n Code Python
- "Why does my Code Tool throw 'Wrong output type returned'?" → Activates: n8n Code Tool
The skills work together seamlessly. For instance, when you ask "Build and validate a webhook to Slack workflow", multiple skills collaborate: Workflow Patterns, MCP Tools Expert, Node Configuration, Code JavaScript, Expression Syntax, and Validation Expert.
Additionally, the pack includes an always‑on enforcement layer (router skill and PreToolUse/PostToolUse hooks) that surfaces the right guidance at the moment of decision. This layer is active only in the Claude Code / Codex plugin installation; on Claude.ai the skills still activate by description but without the proactive nudges.
What You Need to Provide
- n8n-mcp MCP server: Installed and configured (Installation Guide).
- Claude Code, Claude.ai, or Claude API access.
.mcp.jsonconfiguration: Configured with the n8n-mcp server.
What You Will Get
- The ability to build production‑ready n8n workflows with confidence, avoiding common mistakes.
- Correct n8n expression syntax and data mapping.
- Effective use of n8n-mcp MCP tools.
- Proven workflow patterns for different scenarios.
- Proper node configuration and error handling.
- Correct code in JavaScript, Python, and Code Tool nodes.
- Best practices for binary data, sub‑workflows, AI agents, multi‑instance, and self‑hosted deployment.
Usage Examples
The skills activate automatically based on your query. See the examples in the "How It Works" section above.
Important Notes
- Prerequisites: You must have the n8n-mcp MCP server installed and configured, and a compatible AI assistant (Claude Code, Claude.ai, or Claude API).
- Hooks enforcement: The proactive nudges (router skill, PreToolUse/PostToolUse hooks) only run in the Claude Code / Codex plugin installation. On Claude.ai, the skills still activate by description but without the proactive reminders.
- Limitations: The Code Python skill warns that no external libraries (requests, pandas, numpy) are available in the n8n Code node. The Code Tool skill has a different contract than the regular Code node.
- Compatibility: Designed for n8n workflows built with the n8n-mcp MCP server. The self-hosting skill is for on-premise deployment, not for n8n Cloud.