selectalisted

# selecta **This skill starts audio.** Run it only when the user asked for music in their own words. Setting up an environment, starting a long task, or the user saying "I'm about to focus for two hours" or "this refactor is going to be boring" are **not** requests for music. Say nothing about music in those cases. Everything runs through one executable: ``` "${CLAUDE_PLUGIN_ROOT}"/bin/selecta <subcommand> [args] ``` Run it with Bash and report what it prints. Its output is written to be read back to the user almost verbatim. ## What this actually is The status line comes first. selecta shows what is playing **whatever is playing it** — Spotify, Apple Music, or any MPRIS player on Linux — and needs nothing installed to do that. Its own radio playback is a bonus for people who want music without leaving the terminal. So `selecta` with no arguments is a question, not a command. It never starts anything. ## The one rule that matters **Never construct a stream URL, station id or playlist entry yourself.** Resolution belongs to `selecta resolve`, which owns the catalogues and the provider APIs. A URL composed from memory will be wrong, dead, or an invention. If `resolve` returns nothing, say so and offer `selecta stations`. ## Playing ``` selecta play amapiano # a genre selecta play something calm for reading # a mood, in the user's words selecta play # resume what this repo sounds like ``` Pass the user's words through