Generative Media Skills provides AI agents with structured, CLI-powered primitives for multimodal media generation including text-to-image, text-to-video, and text-to-audio via muapi.ai. It exposes a Model Context Protocol server and expert skill recipes that wrap over 100 models (Midjourney, Flux, Kling, Veo3, Suno) into agent-friendly pipelines with JSON output, semantic exit codes, and local file upload support.