eval on libero with joint action space model

[D222097]

Hi, I am trying to add ForwardKinematicsJointsToEEAction to my env_postprocessor steps in order to adapt my joint action space model for evaluation in the libero environment, but the resulting end-effector actions appear to be incorrect.
I added the following code:

from lerobot.robots.so100_follower.robot_kinematic_processor import ForwardKinematicsJointsToEEAction
from lerobot.model.kinematics import RobotKinematics
postprocessor_steps.append(
            ForwardKinematicsJointsToEEAction(
                kinematics=RobotKinematics(
                    urdf_path='/xxx/lerobot/src/lerobot/envs/franka_description/panda.urdf',
                    target_frame_name="panda_link8"
                ),
                motor_names=[
                    "joint_1",
                    "joint_2",
                    "joint_3",
                    "joint_4",
                    "joint_5",
                    "joint_6",
                    "joint_7",
                ],
            )
            )

I am using a URDF file found online and am unsure if it accurately matches the Franka robot used in the libero environment.
I am not certain if there is a recommended alternative approach for converting joint actions to end-effector (EEF) actions in libero.

Is there an official URDF file or recommended configuration for the libero environment?
Any guidance would be appreciated. Thank you in advance for your support.