Sounds great; Citoid is an amazing tool and I'm glad to see more people make use of it.
You can tell which Zotero translator is used by hovering over the Save to Zotero icon. It will save "Save to Zotero" with the translator in parentheses.
In your case, this will all be the Embedded Metadata Translator, a generic translator that's very hard to understand when you start out and a very poor choice for modifying.

Instead there are three options:
1. Use a translator for another Newspaper as a template. The various German newspapers might not be a bad choice (don't pick the NY Times; that one is rather complicated)
2. Build a Framework translator. Several US magazines and newspapers have those, mostly built by me. These have by far the lowest learning curve, all you need essentially is a basic understanding of xpaths and a little bit of regular expressions. No javascript at all.
3. If you find this currently works quite well already, you can call the embedded metadata translator from another translator and then just supplement the data (this isn't possible using the Framework, though).

You should be able to easily find documentation for 2) and examples for 3) (I'm in a rush right now), but let me know if you need more help on this.

Thanks adamsmith and zuphilip. Now I've found the way to detect which translator Zotero uses! Thanks for the compliments to Citoid. I will do my best to pass them to Marielle Volz. There are a couple of people who do use VisualEditor and Citoid in the Netherlands, and they love it. Great work by the Zotero team, and thanks to the creators of all the translators, of which many were contributed by you.

I have played with the Die Zeit and the Le Monde translators. I can get in Scaffold the Run doWeb to work and produce output. So far I'm stuck with "detectWeb did not match". So I need to work on that. Indeed the Embedded Metadata is a complex piece of code.

NU.nl actually has some embedded metadata which I could find in the page source of http://www.nu.nl/buitenland/4175724/meer-rechten-saudische-ex-vrouwen-en-weduwen.html:

var context_dc_params = 'CATEGORY=algemeen;KEYWORDS=saudi-arabië;',
context = {
'dc_params': context_dc_params,
'category' : 'algemeen',
'subcategory' : 'buitenland',
'article_id' : '4175724',
'article_name' : 'meer-rechten-saudische-ex-vrouwen-en-weduwen',
'article_author' : 'nu.nl/ANP'
};

There is something strange happening. A reference created directly from the webpage with Zotero, and exporting from Zotero to Wikipedia Citation Template will yield {{Cite web| title = Meer rechten voor Saudische ex-vrouwen en weduwen| work = NU| accessdate = 2015-12-02| date = 2015-12-02| url = http://www.nu.nl/buitenland/4175724/meer-rechten-saudische-ex-vrouwen-en-weduwen.html}}

However adding the reference using citoid yields.<ref>{{Cite web|title = Meer rechten voor Saudische ex-vrouwen en weduwen|url = http://www.nu.nl/buitenland/4175724/meer-rechten-saudische-ex-vrouwen-en-weduwen.html|website = NU|publisher = https://plus.google.com/+NUnl|accessdate = 2015-12-02|language = nl-NL}}</ref>

The strangeness is in the publisher field to the latter. NU.nl does have a google plus account, but some people don't expect a url here. In the pagesource there is indeed a tag <link href="https://plus.google.com/+NUnl" rel="publisher" />. In Citoid this pops up, in Zotero it doesn't.

This post was just to let know I've read your replies and that I'm trying to understand the tools provided.