This is a surprisingly sophisticated NetEase Cloud Music assistant that runs on model judgment rather than hardcoded rules. It analyzes your liked songs (up to 200) to build a preference profile including temporal patterns (when you favorite tracks, what genres at what times), then uses that data to craft intelligent search strategies across playlists, albums, and singles. The two-layer recommendation explanations are clever, linking specific evidence from your listening history to concrete playlist characteristics. It handles scheduled pushes via cron, can create playlists automatically, and integrates with Feishu for richer notifications with cover images. The content safety checks upfront and the deduplication logic (tracking what's already been recommended) show practical production thinking. Requires ncm-cli for actual playback control, but the real value is in the analysis and curation layer.
npx skills add https://github.com/netease/skills --skill netease-music-assistantjuliusbrussee/caveman
mattpocock/skills
mertbuilds/skills
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills