motion_bench_maker: a possible EVA-CLIENT integration

Hi maintainers,

I came across [KavrakiLab/motion_bench_maker](https://github.com/KavrakiLab/motion_bench_maker) and noticed that it focuses on “Generate and benchmark datasets for robot manipulation problems”. That is close to a problem we are working on with [EVA-CLIENT](https://github.com/Noietch/EVA-CLIENT), a unified framework for real-robot deployment, evaluation, and LeRobot data collection.

If this direction looks relevant, you can take a look at EVA-CLIENT here: https://github.com/Noietch/EVA-CLIENT

For KavrakiLab/motion_bench_maker, EVA could add a real-robot validation layer that replays the project's data format while storing trial-level video, state, and outcome metadata. A small first experiment could be to replay one representative task and compare offline records with one real-robot trial. We would be happy to prepare a minimal adapter or example around your current interfaces, rather than asking you to change the project architecture.

If this is useful, I would be glad to compare notes or help with a small integration. If it is not relevant, no worries—and we will not follow up.

Best,
EVA-CLIENT Team

Start by reviewing the motion_bench_maker repository and the linked EVA-CLIENT framework to understand their current data interfaces. The proposed first experiment is a minimal adapter or example that replays one representative task on a real robot and records trial-level video, state, and outcome metadata. Done means comparing the offline records with one real-robot trial.

Written by the indexing model from the issue text.