FWIW, For those in future interested in an answer for this.
The human finger is 98% water which has a dielectric constant of 80 compared to plastic which is ~ 4. The required test must provide some insulation material which is soft and has this equivalent dielectric constant and make constant, with a surface area just slightly above the minimum required to be detected.
The only dielectrics approaching or exceeding this are toxic materials inside wet capacitors. Glycolic Nitrile gloves used in medicine are about 27 would be perfect to protect glass yet thin enough to couple adequate pF to touch screen yet isolated between the contacts the size of a large straw but may be short in length to simulate a finger making contact.
However the capacitive touch screen must require a bridge capacitance across the matrix grid of the electrodes beneath the thin glass surface. To make contact with the simulated finger is easy but to modulate its local bridge capacitance is not easy without motion.
Perhaps two tiny finger contacts in close proximity can be bridged with an electronic switch to simulate a finger contact but this depends on the algorithm used to avoid false contacts such as rain drops.

The relationship between the size of the contact area to the tiny electrodes or in this case 2 tiny fingerpads close together than can be separated and switched electronically together to become 2 series capacitors shunting the glass surface may be possible but difficult to overcome the touchpads false detection algorithms.

simulate this circuit – Schematic created using CircuitLab