Hey n8n community 
I just published n8n-nodes-agent-kit — a community node package that turns
n8n into a serious platform for building AI agents in production.
It ships 6 nodes that work together:
Agent Kit — single agent with tool calling, memory, skills, guardrails
Orchestrator Kit — supervisor that delegates to specialist subagents,
returns full executionTrace with per-agent token usage
Sub Agent Kit — specialist agent that plugs into the Orchestrator
Agent Memory — SQLite long-term + in-process session buffer
MCP Gateway — speaks Model Context Protocol both as client and server
Skill Loader — loads SKILL.md files inline or from a GitHub repo
(agentskills.io format)
Powered by OpenRouter out of the box, so you get Claude, GPT-4o, Gemini,
Llama, Qwen — 400+ models with one credential.
Why I built it
The built-in AI Agent is great for prototypes, but every production project I
shipped needed: persistent memory, multi-agent orchestration, MCP tools,
observability per agent, and guardrails. I kept gluing the same pieces with
Code nodes — so I packaged them properly.
Install
Settings → Community Nodes → Install → n8n-nodes-agent-kit
Links
npm: https://www.npmjs.com/package/n8n-nodes-agent-kit
GitHub: GitHub - lucasbrito-wdt/n8n-nodes-agent-kit · GitHub
Issues / feedback welcome
Would love to hear what you’d build with it — and PRs / suggestions are very
welcome.