n8n

A setup guide for connecting CEL, an agent execution system, to n8n, a workflow automation platform. It describes a currently usable command-line connection and a planned HTTP connection for longer or remote work.

In plain words
What is it for?
Use it to run CEL goals or actions from n8n workflows through the Execute Command node, or plan a future HTTP Request integration.
Why use it?
It clarifies which system handles scheduling, branching, retries, and execution, and shows how to connect them without assuming n8n supports MCP directly.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/dimpagk92/cellar/n8n
Clone the repo
git clone --depth 1 https://github.com/dimpagk92/cellar
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,640 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.01640
Opus 5 $0.00000 $0.00820
Sonnet 5 $0.00000 $0.00328
Haiku 4.5 $0.00000 $0.00164

Measured 2d ago against content hash f1c1b25517f0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

n8n scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

docs/agents/n8n.md · 149 lines

How it starts

The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Driving CEL from n8n

This page shows how to drive CEL from n8n.

Read docs/adapters-cel-agents.md first if you haven't.

Purpose

n8n is a workflow automation platform. It does not speak MCP natively, so CEL needs a bridge. Two paths:

  1. Execute Command node → cellar CLI. Works today. Each step is a shell invocation. Good for deterministic, short workflows.
  2. HTTP Request node → cellar-worker. Planned. Uses the Phase 1 worker protocol (docs/worker-protocol.md). Good for long-running goals and remote execution. Not yet shipped — tracked on the roadmap.

This document covers both. Path 1 is usable now.

Ownership

  • n8n owns: the workflow graph, branching, error routing, retries, and scheduling.
  • CEL owns: execution of each goal or action.
  • Adapters own: app-specific truth (Numbers cells etc.).

Because n8n is a workflow engine, not an LLM agent, it typically either:

  • (a) hands CEL a full natural-language goal and lets cel_think run_goal drive autonomously, or
  • (b) hands CEL discrete actions produced by some upstream LLM node (e.g. an OpenAI node) and composes them with cel_act.

Mode (a) is the simplest integration and is what the examples below show.

Path 1: Execute Command → cellar CLI (works today)

Setup

  1. Build CEL on the machine running n8n:
    cd /path/to/cellar && pnpm install && pnpm -r build
    
  2. Make sure cellar is on PATH (or reference it by absolute path in the n8n node).
  3. Grant macOS Accessibility permission to whichever process n8n uses to shell out (e.g. node, docker, or the n8n desktop app).
  4. If you use cel_think run_goal from the CLI, set CEL_LLM_PROVIDER, CEL_LLM_API_KEY, and CEL_LLM_MODEL in the n8n environment.

Example n8n Workflow Snippet

Minimal workflow: manual trigger → Execute Command node that runs a CEL goal → Set node that surfaces the result.

{
  "nodes": [
    {
      "parameters": {},
      "id": "trigger",
      "name": "Manual Trigger",
      "type": "n8n-nodes-base.manualTrigger",
      "typeVersion": 1,
      "position": [240, 300]
    },
    {
      "parameters": {
        "command": "cellar run-goal \"Open Numbers and write BTC in A1\" --json"
      },
      "id": "cel-run",
      "name": "CEL Run Goal",
      "type": "n8n-nodes-base.executeCommand",
      "typeVersion": 1,
      "position": [520, 300]
    },
    {
      "parameters": {
        "values": {
          "string": [
            { "name": "result", "value": "={{ $json.stdout }}" }
          ]
        }
      },
      "id": "result",
      "name": "Result",
      "type": "n8n-nodes-base.set",
      "typeVersion": 2,
      "position": [800, 300]
    }
  ],
  "connections": {
    "Manual Trigger": { "main": [[{ "node": "CEL Run Goal", "type": "main", "index": 0 }]] },
    "CEL Run Goal": { "main": [[{ "node": "Result", "type": "main", "index": 0 }]] }
  }
}

Read the full file on GitHub · 149 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 149 lines · 0 tokens per session scan A f1c1b25517f0

Subscribe to this mod's changes

n8n is an agent published in the GitHub repository dimpagk92/cellar (4 stars, last pushed 22d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 1,640 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.