Not sure I agree with this - what good would it do to add this cell-type?
How would it be rendered in the file output?

My use case for adding cell type is I'm attempting to generate XLS' of varied external data. Often enough a column of objects or other data pops up that isn't one of the ExcelJS types, and this gives a fall back output of JSON.stringified() which I think is better than breaking.

Re: How would it be rendered in the file output?

Any nested objects will be stringified.

{
    foo: {
        name: "Rolly",
        id: 12
    },
    baz: {
        property: 'value',
        deep: {
            evendeeper: {
                mariana: "trench"
            }
        }
    }
}

becomes

"{"foo":{"name":"Rolly","id":12},"baz":{"property":"value","deep":{"evendeeper":{"mariana":"trench"}}}}"

We can all agree that the output isn't exactly legible for my grandmother, but I think behaviour should be to output a spreadsheet of the data that is passed in, rather than failing or worse dropping.

The primary scenario is a client saying "hey all that data in the DB, can we have a copy?", and being able to export complete datasets - flattening and normalisation techniques are, by design, manipulating data.