This paper presents a method for enabling manipulation skills using the legs of a quadruped robot, known as pedipulation. The approach involves training a reinforcement learning policy to track foot position targets, allowing the robot to perform a variety of real-world tasks such as door opening, sample collection, and obstacle pushing. The controller is designed to be robust to disturbances, has a large workspace through whole-body behaviors, and can reach far-away targets with gait emergence, enabling locopedipulation. The controller was deployed on a quadrupedal robot using teleoperation, demonstrating various real-world tasks. The controller is robust to interaction forces at the foot, disturbances at the base, and slippery contact surfaces. The controller was tested in simulation and hardware experiments, and it was shown to be effective in a range of manipulation tasks. The controller was also tested on irregular terrain to ensure robustness. The controller was trained using a curriculum-based command sampling approach, allowing the robot to adapt its stance and locomote towards far-away targets. The controller was also tested for its ability to handle disturbances and was shown to be robust to external forces. The controller was able to carry more than 2.0 kg at the foot and was able to press a button far above the robot's base. The controller was also able to move obstacles out of the way. The controller was shown to be effective in a range of manipulation tasks, including door opening, rock sample collection, and pushing obstacles. The controller was also shown to be effective in a range of real-world scenarios, including maintenance, home support, and sample collection. The controller was able to handle a variety of tasks without task-specific adaptations. The controller was also shown to be effective in a range of environments, including wet surfaces and slippery contact surfaces. The controller was able to handle a variety of tasks, including pushing objects, opening doors, and collecting samples. The controller was also shown to be effective in a range of scenarios, including planetary exploration, where the robot needs to have minimal mass and mechanical complexity. The controller was able to handle a variety of tasks, including pushing objects, opening doors, and collecting samples. The controller was also shown to be effective in a range of environments, including wet surfaces and slippery contact surfaces. The controller was able to handle a variety of tasks, including pushing objects, opening doors, and collecting samples. The controller was also shown to be effective in a range of scenarios, including planetary exploration, where the robot needs to have minimal mass and mechanical complexity.This paper presents a method for enabling manipulation skills using the legs of a quadruped robot, known as pedipulation. The approach involves training a reinforcement learning policy to track foot position targets, allowing the robot to perform a variety of real-world tasks such as door opening, sample collection, and obstacle pushing. The controller is designed to be robust to disturbances, has a large workspace through whole-body behaviors, and can reach far-away targets with gait emergence, enabling locopedipulation. The controller was deployed on a quadrupedal robot using teleoperation, demonstrating various real-world tasks. The controller is robust to interaction forces at the foot, disturbances at the base, and slippery contact surfaces. The controller was tested in simulation and hardware experiments, and it was shown to be effective in a range of manipulation tasks. The controller was also tested on irregular terrain to ensure robustness. The controller was trained using a curriculum-based command sampling approach, allowing the robot to adapt its stance and locomote towards far-away targets. The controller was also tested for its ability to handle disturbances and was shown to be robust to external forces. The controller was able to carry more than 2.0 kg at the foot and was able to press a button far above the robot's base. The controller was also able to move obstacles out of the way. The controller was shown to be effective in a range of manipulation tasks, including door opening, rock sample collection, and pushing obstacles. The controller was also shown to be effective in a range of real-world scenarios, including maintenance, home support, and sample collection. The controller was able to handle a variety of tasks without task-specific adaptations. The controller was also shown to be effective in a range of environments, including wet surfaces and slippery contact surfaces. The controller was able to handle a variety of tasks, including pushing objects, opening doors, and collecting samples. The controller was also shown to be effective in a range of scenarios, including planetary exploration, where the robot needs to have minimal mass and mechanical complexity. The controller was able to handle a variety of tasks, including pushing objects, opening doors, and collecting samples. The controller was also shown to be effective in a range of environments, including wet surfaces and slippery contact surfaces. The controller was able to handle a variety of tasks, including pushing objects, opening doors, and collecting samples. The controller was also shown to be effective in a range of scenarios, including planetary exploration, where the robot needs to have minimal mass and mechanical complexity.