in neuralforcast library for what we can use static feature and dynamics df and also new data can any one told me the difference between those parameters in ml forecast because i need to add feature not just the lag and do i need to add the lag feature to predict in ml forecast models or not

Static features are features that dont change over time. Something like category of a product. Dynamic features are things like the price of the item or temperature at that day. Here is a tutorial for neural: https://nixtla.github.io/neuralforecast/examples/exogenous_variables.html