43 Sunsets · AI agent skills · documentationn8n Production-Readiness Auditor
Paste an n8n workflow export, get a graded 0–20 report across 10 categories with exact node-level fixes. Static analysis only — the skill never runs your workflow or calls your services.
Get it on Agensi — $12One-time purchase · works with SKILL.md-compatible agents (Claude Code, Cursor and others — the Agensi page shows the current list) · free lite version (3 checks) if you want to try the idea first.
How to use it 3 steps
- In n8n, open the workflow → menu (⋯) → Download. This saves the workflow as a JSON file. Credential secrets are not included in the export (only credential names).
- Give your agent the skill (Agensi shows the install step for your agent), then paste the JSON and ask:
Audit this n8n workflow export for production readiness. - Read the report top-down: Critical → Important → Nice-to-have, then the 15-minute fix plan. Apply the fixes in a copy of the workflow first, run it once, then promote.
What the report contains
- Overall score 0–20 with a verdict (e.g. NEEDS WORK, PRODUCTION READY). A solid workflow is told it is solid — the skill does not invent findings.
- Findings by severity — each names the exact node, the exact setting and the exact value to set. Never “add error handling”; always something like
retryOnFail=true, maxTries=3, waitBetweenTries=5000on a named node. - Blast radius in business terms — what actually goes wrong for your customers if the finding is left as is.
- A 15-minute fix plan, ordered by damage prevented.
Illustrative excerpt — this is the report shape, not output from your workflow:
# Production Readiness Report Overall: 12/20 — NEEDS WORK ## Critical - [C1] Hardcoded API key in 'HTTP Request' → Fix: remove the key from the 'Authorization' header parameter; use a Header Auth credential (name: 'HTTP Header Auth', header 'Authorization'). - [C2] No Error Workflow → Fix: Create a global error handler and set settings.errorWorkflow.
The 10 categories
| # | Category | What it looks for |
|---|---|---|
| 1 | Error handling | Error workflows, retries, failure notifications |
| 2 | Credential hygiene | Hard-coded secrets and API keys — flagged CRITICAL |
| 3 | Idempotency | Will a double-fired trigger send duplicate emails or charges? |
| 4 | Rate limits & batching | Worst-case API cost estimated in dollars and calls |
| 5 | Webhook security | Open webhooks that mutate data — flagged CRITICAL |
| 6 | Timeouts | Long-running paths without limits |
| 7 | Observability | Will you know it worked — not just that it failed? |
| 8 | Data handling | What PII leaves your tenant, and how to minimize it |
| 9 | Naming & maintainability | Because 3 am debugging is real |
| 10 | Cost traps | Tight polling, oversized LLM context, infinite-loop risk |
What it does not do
- It does not execute the workflow, call the services it references, or fetch anything. Your JSON stays in your conversation.
- It cannot detect runtime-only bugs (wrong data at runtime, API auth failures, upstream outages). It reviews the exported configuration.
- It needs the full JSON export — screenshots or partial snippets are not enough.
- It is a technical review, not a guarantee. Test changes in a non-production copy before deploying.
FAQ
Does it work with self-hosted n8n as well as n8n Cloud?
Yes. The audit reads the exported workflow JSON, which has the same shape on n8n Cloud and self-hosted installs.
Is my workflow data private?
The skill is a set of instructions for the agent you already use; nothing is sent to 43 Sunsets. What leaves your machine is governed by your agent’s own privacy settings.
How is this different from asking a general AI to “review my workflow”?
The skill fixes the checklist (10 categories), the output shape (score, severity, exact node/setting/value, blast radius, fix plan) and the honesty rule (no invented findings). A general prompt gives you general advice.
Refunds?
Purchases are covered by Agensi’s 30-day refund policy under Agensi’s terms; refunds are handled by Agensi, not by 43 Sunsets.
Also by 43 Sunsets: API Breaking-Change Detector · all products · [email protected]
Not affiliated with, endorsed by, or sponsored by n8n GmbH. “n8n” is a trademark of its respective owner, used here only to describe compatibility.