CARLA
Official@carla-simulator
CARLA is an open-source autonomous driving simulator with skills for scenario execution, sensor simulation, ROS 2 integration, and Unreal Engine builds.
Agent Skills by CARLA
Showing 60 vetted skills indexed across 1 GitHub repositories.
carla-simulatorcontrol-traffic-lights
Controls CARLA traffic-light actors to set states, freeze phases, and adjust timing.
carla-simulatorcontrol-lights
Controls CARLA map light sources via the LightManager API.
carla-simulatorload-map
Loads, reloads, and reshapes maps on a running CARLA server via the Python API.
carla-simulatordebug-draw
Draws debug overlay primitives in a running CARLA world via the world.debug Python API.
carla-simulatorcontrol-vehicle
Drives a CARLA vehicle directly via VehicleControl, Ackermann speed targets, lights, doors, and physics tuning.
carla-simulatorbounding-boxes
Extracts 3D actor bounding boxes from CARLA and projects them into camera images as 2D annotations.
carla-simulatorset-weather
Reads and sets weather parameters on a running CARLA server via the Python API.
carla-simulatorspawn-vehicles
Spawns and destroys autopilot vehicles on a running CARLA server via the Python API.
carla-simulatorcontrol-traffic
Configures CARLA Traffic Manager settings for autopilot vehicles globally or per vehicle.
carla-simulatorreplay-recording
Replays CARLA recorder .log files with time windows, speed control, and actor following.
carla-simulatorset-world-settings
Reads and changes a running CARLA server's WorldSettings including synchronous mode, fixed timestep, and rendering.
carla-simulatorspawn-walkers
Spawns and destroys autonomous wandering pedestrians in a CARLA simulation via the Python API.
carla-simulatorread-sensor
Captures, displays, and summarizes data streams from CARLA simulator sensors.
carla-simulatortoggle-env-objects
Lists and toggles CARLA map environment objects by CityObjectLabel via the Python API.
carla-simulatorworld-data
Queries the live CARLA world to resolve actor descriptions into concrete actor ids.
carla-simulatordebug-navmesh
Validates and visualizes a CARLA map's pedestrian navigation mesh by sampling walkable locations.
carla-simulatorcreate-sensor
Spawns and attaches cameras, lidar, radar, and other sensors to actors on a running CARLA server.
carla-simulatormap-waypoints
Queries and visualizes a CARLA map's road network through the Map and Waypoint API.
carla-simulatortelemetry
Reads frame-consistent live telemetry from a CARLA actor via the Python API.
carla-simulatorrecord-simulation
Records a running CARLA simulation to a verified .log file for later replay.
carla-simulatornavigate-to
Drives a CARLA vehicle to a destination using navigation agents or plans routes with GlobalRoutePlanner.
carla-simulatorquery-recording
Inspect CARLA .log recordings for header info, collisions, and blocked actors without replaying them.
carla-simulatorcontrol-spectator
Moves, frames, and follows actors with the CARLA spectator camera via the Python API.
carla-simulatorrun-scenic-scenario
Runs Scenic probabilistic scenarios against a CARLA server with bounded, verified simulations.
Frequently Asked Questions About CARLA
FAQPage SchemaWhat tasks can I perform with CARLA's skill collection?▼
You can launch servers, load maps and OpenDRIVE worlds, spawn vehicles and pedestrians, control traffic lights and weather, attach cameras/lidar/radar/IMU sensors, record and replay simulations, run Scenic and OpenSCENARIO scenarios, evaluate agents on the Leaderboard, and build or package CARLA on Unreal Engine 4.26 and 5.8.
Who is CARLA's skill set designed for?▼
It targets autonomous driving researchers, perception and planning engineers, and simulation developers who need synthetic sensor datasets, scenario-based validation, Leaderboard benchmarking, ROS 2/Autoware integration, or custom map, vehicle, walker, and prop assets inside a CARLA simulator.
How do I install and run CARLA with these skills?▼
Use download-carla for releases, nightly builds, source checkouts, or Docker images, then install-python-api to add the matching carla client wheel. Most runtime skills need only a reachable running server; source builds on UE4 or UE 5.8 require Linux, CMake, and 120-400 GB of disk.
Is CARLA free and what license applies?▼
Yes. Every skill in this manifest is released under the MIT license, and CARLA itself is open source. Costs are only infrastructural: releases are 8-10 GB downloads, source builds need an Epic-linked GitHub account for the Unreal fork, and rendering servers require an NVIDIA GPU.
What prerequisites do CARLA skills require?▼
Runtime skills need the CARLA PythonAPI installed and an already-running server, tested against 0.9.16 and 0.10.0. ScenarioRunner skills need a matching branch checkout; Leaderboard 2.x needs Town12/Town13 maps; ROS 2 features need a server built with -DENABLE_ROS2=ON; UE 5.8 builds need CMake 3.28+, ninja, and git-lfs.