What problem does it solve? Tuning how autopilot traffic behaves in a CARLA simulation requires knowing dozens of Traffic Manager API calls, their sign conventions, and sync-mode coupling. This Skill encodes those procedures so an agent can change traffic speed, following distance, light-running, lane changes, and determinism without improvising against the Python API. ## Core Features & Use Cases - Global traffic tuning: Set speed difference, following distance, random seed, hybrid physics, dormant-vehicle respawn, and OSM mode for every vehicle on a Traffic Manager port. - Per-vehicle control: Target one car by id, role, or filter to adjust its speed, lane offset, light/sign/vehicle ignore percentages, lane-change permission, and TM-managed lights. - Deterministic and synchronous runs: Seed the TM and enable TM synchronous mode coordinated with the world's sync mode for reproducible simulations. - Use Case: A user says "make the traffic move faster and keep more distance" — the Skill runs traffic.py global --speed-diff -30 --distance 3 against the running server after verifying prerequisites. ## Quick Start Ask the agent to make the autopilot traffic in your running CARLA server faster or slower, and it will check prerequisites and apply the Traffic Manager settings for you.
