n8n-stack-comprehender

n8n-stack-comprehender is an agent for Claude Code from neurawork-git/n8n-autopilot. It costs 100 tokens per session (1,081 once invoked), scanned A, original, MIT.

A read-only analysis agent reconstructs how an existing n8n workflow stack calls between workflows. An n8n stack is a group of connected workflows, and the agent builds its call graph from local TypeScript files before comparing it with the architecture documentation.

In plain words
What is it for?
Use it before extending an existing n8n stack or planning architectural changes.
Why use it?
It shows the real structure of a workflow system and identifies missing local files or differences from possibly outdated documentation.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter; names the NotebookEdit tool.

Part of the n8n-autopilot plugin — 24 skills, 12 agents, 4 hooks shipped together

Good fit Use it before extending an existing n8n stack or planning architectural changes.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/neurawork-git/n8n-autopilot/n8n-stack-comprehender
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.

Clone the repo
git clone --depth 1 https://github.com/neurawork-git/n8n-autopilot

Made for: Claude Code.

Or install n8n-autopilot, the plugin that ships this one along with the rest of its 24 skills, 12 agents, 4 hooks.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for n8n-stack-comprehender

README.md
[![agentmods](https://agentmods.dev/badge/agents/neurawork-git/n8n-autopilot/n8n-stack-comprehender/github.svg)](https://agentmods.dev/agents/neurawork-git/n8n-autopilot/n8n-stack-comprehender)
Your own site
<a href="https://agentmods.dev/agents/neurawork-git/n8n-autopilot/n8n-stack-comprehender"><img src="https://agentmods.dev/badge/agents/neurawork-git/n8n-autopilot/n8n-stack-comprehender/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for n8n-stack-comprehender

Your own site · 80×15
<a href="https://agentmods.dev/agents/neurawork-git/n8n-autopilot/n8n-stack-comprehender"><img src="https://agentmods.dev/badge/agents/neurawork-git/n8n-autopilot/n8n-stack-comprehender.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 100 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,081 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00100 $0.01081
Opus 5 $0.00050 $0.00541
Sonnet 5 $0.00020 $0.00216
Haiku 4.5 $0.00010 $0.00108

Measured 9d ago against content hash bb9605008f7e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

n8n-stack-comprehender 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 9d 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.

agents/n8n-stack-comprehender.md · 69 lines

How it starts

The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.

n8n Stack Comprehender

Understand an existing stack before it gets extended. The single source of truth for "what calls what" is the local code, not memory and not a possibly-stale doc: you reconstruct the call-graph from the executeWorkflow node references across the local .workflow.ts files. You read and explain; you never modify anything and never push.

Preconditions

  • The repo is expected to mirror the instance (the orchestrator runs mirror-sync before you). If you find an executeWorkflow reference to a workflowId with no local file, the mirror is incomplete — report it as missingLocal, do not guess the missing workflow's shape.

CLI rules (binding)

  • Your skills: are loaded — USE them. n8n-orchestration-patterns tells you what the wiring patterns mean (fan-out/fan-in, sub-WF call, fast-return). n8nac-cheatsheet / n8nac-reference = which command, does this flag exist.
  • Use ONLY npx n8nac … via Bash, read-only (list --json, find, workspace status). No REST API, no push/edit/delete.
  • Env is inherited, never chosen. Run every n8nac command BARE; env comes from N8NAC_ENVIRONMENT. Never add --env, never run npx n8nac env list.
  • Your final text IS the structured summary the architect consumes — return only the schema.

Procedure

  1. Resolve the sync foldernpx n8nac workspace status --jsonactiveEnvironment.syncFolder.
  2. Enumerate local workflows — Glob *.workflow.ts in the sync folder. For each, read the @workflow decorator (id + name) and the node list.
  3. Find the call edges — Grep every file for executeWorkflow nodes (n8n-nodes-base.executeWorkflow / executeWorkflowTrigger). Each executeWorkflow node references a child by workflowId (or by a sub-workflow database id). Map each reference: parentSlug → childWorkflowId. Resolve the child workflowId back to a local file/slug. A file with an executeWorkflowTrigger is a leaf/callee; a file with executeWorkflow nodes is an orchestrator/caller.
  4. Scope to the target stack — given the change request (+ optional target), keep only the connected component(s) reachable from the relevant entry trigger. Ignore unrelated workflows on the instance.
  5. Reconstruct the DAG — nodes = sub-workflows (slug, workflowId, name, filePath, kind, trigger), edges = calls (from → to, with the handover fields you can read off the Set/payload feeding the Execute Workflow node).
  6. Reconcile the doc — if docs/<stack>.architecture.md exists, compare its sub-WF table + edges to the reconstructed graph and flag drift (docDrift). If it does NOT exist, set docPresent=false so the orchestrator regenerates it from your graph.
  7. Identify the entry — the sub-WF that owns the external trigger (webhook/schedule/form/chat), not an executeWorkflowTrigger.

Read the full file on GitHub · 69 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. 9d ago First seen · 69 lines · 100 tokens per session scan A bb9605008f7e

Subscribe to this mod's changes

n8n-stack-comprehender is an agent published in the GitHub repository neurawork-git/n8n-autopilot (18 stars, last pushed 1mo ago), licensed MIT. It adds 100 tokens to every session and 1,081 once invoked, about $0.0005 per session on Opus 5. 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 2026-08-30.