Microscopic contributions to the deviation from Amontons friction law

This project is a scientific research artifact rather than a software product. Its value lies in the specific application of Machine Learning Force Fields (MLFFs) to describe friction mechanisms that deviate from classical Amontons's law at the nanoscale. While the 5 forks in 9 days indicate strong interest within a specific research circle, the project lacks a software 'moat.' The defensibility is low because the techniques (MD with MLFFs) are increasingly standard in computational materials science. The primary 'competitors' are other academic groups using frameworks like NequIP, MACE, or DeepMD. Frontier labs (OpenAI/Google) are building foundational models for chemistry (e.g., GNoME, MatterSim), which provide the tools to do this research but are unlikely to target the specific niche of MX2/metal friction. The displacement horizon is relatively short because MLFF architectures are evolving rapidly, and this specific implementation will likely be superseded by more general-purpose foundational materials models within 1-2 years.