n8n-agents

What problem does it solve?

This Skill helps you design and operate n8n AI Agent workflows that reliably choose tools, produce correct structured outputs, handle chat and memory correctly, and avoid common orchestration failures.

Core Features & Use Cases

• Choose the right n8n LangChain node: use Agent for multi-turn reasoning/tool calling and specialized nodes (like chains, classifiers, extractors, and sentiment/summarization) for simpler one-shot or routed tasks.
• Tool-first, prompt-safe behavior: craft tool names and descriptions as modular prompt fragments so the model selects tools correctly and you can debug misbehavior.
• Structured output you can trust: parse JSON with a schema and enable auto-fixing using a coding-capable fixer model to recover from malformed outputs.
• Correct chat/memory/session patterns: wire session IDs consistently into memory and avoid conversation bleed across threads and channels.
• Binary/file boundary guidance: support vision inputs to the model, but avoid passing binary directly into tools; instead use storage-key patterns for agent-tool workflows and use native media nodes for one-shot generation.
• Human review for high-impact actions: wrap destructive or user-visible tools with HITL review nodes and approve the literal parameters the agent will send.

Quick Start

Tell your AI to build an n8n workflow using a Chat Trigger and the Agent node, add memory keyed by sessionId, wire tools with precise names/descriptions, and ensure any structured output is validated with an autoFix structured output parser.