The Pons website is based on the published print dictionaries, isn't it? Surely that's a notch above the sites that only exist online.
For the rest, I wouldn't say never use them at all. Every so often that's exactly where people first come across a possible translation for a less common word, e.g., Siehe auch: sinnreich
But it might help to state explicitly that websites are not as reliable as published dictionaries, so they rank lower as potential sources. I still think it might be useful to rank types of sources in the New Entry instructions, e.g.,
1) monolingual dictionaries from reputable publishers (Duden, DWDS, Oxford, Cambridge, Merriam-Webster, American Heritage, Random House ...)
2) bilingual dictionaries from reputable publishers (Pons/Collins/HarperCollins, Oxford-Duden, Langenscheidt, Muret-Sanders, Cassell's ...)
3) full descriptions with definitions from reputable internet reference sites (specialist glossaries: medical, legal ...; species databases: botanical, zoological ...; encyclopedias: Britannica, Wikipedia ...)
4) full text examples from reputable internet corpuses or archives (text corpuses: Uni-Leipzig, Linguee ...; newspaper and magazine archives; academic publications; university websites; G**gle Books; ...)
5) internet-only word lists and 'dictionaries' (dict.cc, TU-Chemnitz, freedictionary.com, webstersdictionary.org ...)
6) other internet examples
Then you only need a few specific rules:
1) never use only one source in New Entry;
2) be sure to cite sources for both German and English;
3) the lower your sources are on the list, the more different sources you need; and
4) if you first come across a word in a lower-ranking source, it's often possible to confirm it by citing several other corroborating sources.