[Fix] play.py cannot override agent_cfg from command line argument

[Shi-Soul]

Problem
After Hydra integration to play.py, the agent configuration is already built from the entry-point yaml and updated with any command-line overrides.
However, the legacy call to parse_rsl_rl_cfg is still executed inside main, which discards the Hydra-supplied config and rebuilds it from scratch, losing all overrides.

Fix
Replace the redundant parse_rsl_rl_cfg with update_rsl_rl_cfg, which only patches the already-constructed config with CLI flags that are not managed by Hydra (e.g. checkpoint path).
This guarantees that Hydra’s command-line overrides are preserved while keeping the extra CLI tweaks intact.

Original Code:

@hydra_task_config(args_cli.task, "rsl_rl_cfg_entry_point")
def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agent_cfg: RslRlOnPolicyRunnerCfg):
    """Play with RSL-RL agent."""
    agent_cfg: RslRlOnPolicyRunnerCfg = cli_args.parse_rsl_rl_cfg(args_cli.task, args_cli)

Corrected Code:

@hydra_task_config(args_cli.task, "rsl_rl_cfg_entry_point")
def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agent_cfg: RslRlOnPolicyRunnerCfg):
    """Play with RSL-RL agent."""
    agent_cfg = cli_args.update_rsl_rl_cfg(agent_cfg, args_cli)