I am applying for a job and they ask my h-index. The problem is that Scopus and Web of Science list different citation number for my papers and this results in a different h-index. Who should I trust?
The best thing to do would be checking directly why the two numbers differ. Both sites offer detailed information on how many citations got each of of your publications.
It is very likely that a few publications (or citations to them) are simply missing from one of the sites, so this means that the higher result is the correct one. But it could also be an error in the other direction coming from a citation mistakenly considered, or something that should not be a publication. (Note that sometimes Scopus and WOS have different ideas of what counts as "publications", especially when it comes to conference proceedings).
When you apply for the job, specify the source of your H-index as well, so that your information is easy to verify.
If you conclude that one of the two websites is missing some of your publications or citations, you can contact them and ask for a correction. This correction may take weeks to be published, if it is accepted. Since your username looks like an Italian name, I should add that if you are looking for academic jobs in Italy this is an important step to do, because citations on Scopus and WOS are widely used there for bibliometric evaluations (and the abilitazione scientifica nazionale in particular).
Both (and neither).
Like other citation metrics, the h-index compiled by these databases only depend on what's already in those databases. Both of them have quality barriers, which means you could well have been cited somewhere else but neither of these indexed it. That's why your Web of Science h-index is almost certainly going to be lower than your Google Scholar h-index. In the same way I'd guess that your Web of Science h-index is lower than your Scopus one.
In other words, your h-index depends on what database is used to generate it. You could say your h-index is 15 and it wouldn't carry meaning; you need to say your h-index is 15 as measured by Web of Science.
I'm not sure how seriously the job you describe treats the h-index. If given the option, I'd report both of them (and the one by Google Scholar). Otherwise just give the higher number, and expect to defend it in the interview.