Agent task: n8n Gmail workflow with a dry-run switch
What it tests: Turning business rules into an importable n8n workflow with a classifier, branching, a review label for paid offers, placeholder credentials and a dry-run switch that is on by default.
Prompt length: 2,731 characters, roughly 780 tokens (estimate).
Du befindest dich in einem Ordner, der zu meiner lokalen n8n-Installation gehört. Aufgabe: Erstelle einen importierbaren n8n-Workflow als JSON-Datei. Der Workflow soll Folgendes machen: Immer wenn eine neue E-Mail über Gmail …
The source has no licence that allows copying, so only a short excerpt is shown here. The full prompt is one click away at the source.
How to grade it
Valid importable JSON with the requested filename; Gmail trigger, an LLM or classifier step, If/Switch, Gmail reply, Gmail label; mails mentioning budget, paid, long-term, sponsored video or an amount get the 'Sponsoring prüfen' label instead of an auto-reply; dry run is on by default and only logs; no invented credentials; the summary answers all five questions.
Scorecard tick what the model got right: 0/6
Runs in your browser; nothing is saved or sent.
Where models usually fail
Dry run missing or off by default; auto-rejecting every sponsoring mail, paid ones included; fabricated credential IDs; node types that don't exist in n8n.
Running it
Run it one-shot and judge the first answer. Open the result exactly as delivered (double-click for HTML). Give the model a large output budget: long specs get cut off on small context settings.
Source
- Author
- Arnie (GitHub @Arnie936)
- From
- Arnie936 / llm-prompting-tests
- Exact file or post
- https://github.com/Arnie936/llm-prompting-tests/blob/03445cf074852ec5535ece9d123bdff442e63ef5/README.md#12-n8n-gmail-sponsoring-workflow-test
- Licence
- No license file (all rights reserved)
- How we use it
- Excerpt only: open the full prompt at the source
Testing a local model? Check first whether it fits your hardware with Can I run this LLM?