Search for a command to run...
ProactiveBench: Benchmarking Proactiveness in Multimodal Large Language Models