Rodent Model from `dm_control` & Compatibility of MJX

Hi team,

I'm a researcher interested in using the rodent model from [dm_control/locomotion/walker](https://github.com/google-deepmind/dm_control/blob/main/dm_control/locomotion/walkers/assets/rodent.xml) for simulations in MuJoCo, ideally in MJX format. Could this model be officially supported and optimized for use in MuJoCo and potentially MJX?

To remove tendons and adapt to MJX, we've added the following constraint to each joint: `solimplimit="0 .99 .01"`. We based this on the humanoid model in MJX. However, we're not confident this is the correct approach. Could you please provide guidance on best practices for accurately converting models between formats, particularly for the rodent model? Any specific recommendations regarding joint constraints in MJX would be incredibly helpful.

Thanks,
Scott

Start by reviewing dm_control/locomotion/walkers/assets/rodent.xml and comparing its constraints with the humanoid model in MJX. Determine whether official rodent support and a reliable conversion approach are feasible; done would require an agreed compatibility plan and specific guidance for joint constraints.

Written by the indexing model from the issue text.