Uncertainty of Zero

As far as I am aware, per 17025, uncertainties used on certificates requires a calibration uncertainty contributor in the budget which proves traceability. I was told by a coworker who has been at other accredited labs that uncertainties can be generated for zero using a short when testing DMMs. I have heard that there is a lab that a 4-wire short can be sent to get a traceable accredited cal, but otherwise the train of traceability has been lost. I have seen Fluke Park list a copper short on their scope of accreditation, but my information stops there. Does anyone have any experience with uncertainties about a 0 Ohm, 0 V, 0 Amp measurement, or a resource like a NIST publication, or other white paper that talked about it? Thanks.