Aluminum forms a brittle intermetallic with iron (as well as some other elements like nickel and cobalt), which I think is the cause of steel screws seizing in aluminum when unlubricated. Perhaps you can exploit this property to mark visible lines on iron and steel using an aluminum rod. I think there are three fundamental obstacles: surface preparation, activation energy, and aluminum cohesion.
If the aluminum is in contact with grease or iron oxide rather than raw iron, it won’t be able to react. I think that if the surface is reasonably clean to begin with, it should be possible to surround the aluminum “pencil lead” with a friable abrasive “wood” to adsorb the grease and grind through the oxide, maybe a mixture of things like talc, kaolin, aluminum oxide, and binders.
The activation energy problem is that the aluminum needs to be at a relatively high temperature to get this reaction. Since it happens by accident with screws, I think this is a feasible thing to achieve by hand, but it might help if the aluminum is in smallish particles, a foam or sponge, or thin aluminum-foil-like layers, so it won’t conduct the heat away from the reaction surface. Maybe thin fibers of something very hard inside the aluminum would also help by scratching hot particles off the iron.
The aluminum cohesion problem is that we want the newly formed intermetallic to stick to the steel rather than the aluminum. For this we want to keep the “pencil lead” from having too much structural strength of its own; if the aluminum is in the form of, again, small particles, foils, or foam, that should help with this. This is somewhat in conflict with the desire to apply a great deal of pressure to the surface in order to overcome the activation-energy barrier.