And I didn't want to alter the crossbolt safety as it reliably serves the intended purpose. I wanted to leave the rifle as is except eliminate the misfire problem, if possible.

There were many references to an older Rifle Magazine article by Dave Scovill where he details a procedure to correct misfires with a M1886 which has a slightly different mechanism but works on the same rebound principle as the M94.  I remember reading the article and probably still have it in a box somewhere around here.  There are top and bottom prongs on the hammer spring strut that push the hammer forward and then start to pull it back before actually firing to the safe position after firing . He adjusted the balance of the push and pull by removing a bit of the bottom strut (about .030 inch ) which solved his misfire problem.

The M94 has two top and bottom fingers on the mainspring strut that contact upper and lower notches on the hammer.  Removing the tang screw and then the hammer retention screw allows the lower tang to be pulled out
exposing the the mainspring strut and hammer.  After removing the mainspring and hammer the strut and hammer should be checked for burrs that can impede the hammer force in addition to the imbalance of the strut fingers and stone smooth as necessary.  My strut was quite rough along the edges that would contact the mainspring.

There is a convenient shelf on the back of the strut that allows the lower finger length to be accurately measured with a caliper. Mine was .6235 and I stoned both evenly at the same time down to .600 keeping the ends rounded as they started out. Some had reported removing  .035 to .037 and not being able to push the crossbolt safety into place without thumbing the hammer back a bit - I stopped before that point and the crossbolt still works as intended.