Describing and modelling stream/shell-crossing in dark matter clustering is a long-lasting theoretical challenge in large-scale structure formation. Current perturbative approaches, based on effective field theories, address the issue by introducing free parameters (counterterms) that capture these non-linear effects on large scales. In the talk, I will present a Lagrangian formalism that incorporates these effects from first principles, offering a more fundamental perspective. I will discuss some of the recent results and explore future prospects and potential development directions.
