This is the gym goal-based robotic environment wrapper designed for rllab.
Since rllab does not support goal-based environment, if you want to implement RL algorithms under rllab to perform gym goal-based robotic tasks, you'll need to wrap corresponding gym environments to be rllab-based, although rllab is no longer under active development and is maintained under garage.

from gym_env import GoalEnv
from env_utils import normalize_goal_env

env = normalize_goal_env(GoalEnv(env_name))

• env_name is the gym Robotics environment name.
• env is a rllab-based environment.

and it should be good to go.