What problem does it solve? When running several @boost task functions in a single funboost script, users often mistakenly wrap consume() in threads or subprocesses. This Skill clarifies that consume() is natively non-blocking and shows the correct pattern for launching multiple queue consumers sequentially in one process. ## Core Features & Use Cases - Sequential multi-consumer startup: Call func1.consume() and func2.consume() directly in order, with each consumer keeping its own concurrency pool, QPS limit, and retry configuration. - Anti-pattern prevention: Explicitly forbids wrapping consume() in threading.Thread, subprocess, or multiprocessing.Process. - Main-thread keep-alive: Uses enable_ctrl_c_quit_on_windows() so the script stays alive and can be stopped with Ctrl+C. - Use Case: You have two queues (queue_a and queue_b) backed by SQLite and want both consumed in one script; push messages, call consume() on each function in sequence, and both consumers run independently. ## Quick Start Ask the AI to write a funboost script that starts consumers for two @boost task functions in sequence without using threads, including main-thread keep-alive.