What problem does it solve? Configuring CARLA sensors requires knowing blueprint IDs, attribute names, mount transforms, and the non-obvious requirement that a sensor only ticks when something listens to it. This Skill encodes those failure modes so an agent can spawn a correctly configured sensor on a running CARLA server instead of improvising against the Python API. ## Core Features & Use Cases - Sensor spawning and attachment: Spawn cameras (rgb/depth/semantic/instance/optical-flow/normals), lidar, radar, IMU, GNSS, and collision/lane-invasion/obstacle sensors, attached to the ego or any actor with a mount transform and blueprint attributes. - Native ROS 2 publishing: Enable enable_for_ros with --ros, set topic and TF frame naming via --ros-name/--ros-frame-id, and print the exact topics the sensor will publish. - Prerequisite checking and lifecycle: Verify the carla module and server reachability before acting, list available blueprints, and destroy sensors with verified cleanup. - Use Case: A user asks to "put a dashcam on the ego". The Skill checks the server, spawns an 800x600 camera.rgb attached to the hero vehicle at the default mount, prints the sensor id, and hands off to the read-sensor skill for data capture. ## Quick Start Ask the agent to add a camera or lidar to the ego vehicle on the running CARLA server, for example "put an 800x600 dashcam on the ego and publish it to ROS".
