music-video: visuals written to a track's structure, your own audio laid on at the end

The user has a piece of music. This skill produces picture for it: the track's length, tempo, mood and section boundaries are read locally, the visuals are written to that structure, the segments are generated so they stay continuous with each other, and the user's own audio file is muxed on at the end — locally, with no API call and at no cost.

ffmpeg is a hard requirement here, not an enhancement: the last step of the deliverable is a local mux, and stitching the segments needs it too.

This skill is a thin, scenario-specific layer over ofox-video-core. It owns the track analysis, the section-to-segment mapping, the visual prompt craft for music, the multi-job cost arithmetic and the final mux; ofox-video-core owns talking to the Ofox API correctly and safely (the OFOX_API_KEY handling, the no-resubmit rule, error-code mapping, download/verification, and reporting the downloaded file's absolute path). Read that skill's safety contract before using this one — it is not restated here.