@maschinenlesbar.org/fim-portal-cli
    Preparing search index...

    Variable SearchCsvResourceValuesConst

    SearchCsvResourceValues: readonly [
        "schema",
        "document-profile",
        "field",
        "group",
        "leistung-steckbriefe",
        "processclass",
        "process",
    ] = ...

    The resource values tools/search-csv-download recognises — the resource names of the portal's own search page (fimportal.de/search?resource=...). The OpenAPI spec types the parameter as a free string, and the server never rejects a value: anything it does not recognise (e.g. the plural schemas) silently exports Leistungen instead. So tools.searchCsvDownload requires one of these and throws FimValidationError otherwise (searchCsvResourceProblem); the CLI's --resource choices are this list.