[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M71HVB9501BZKWRX5M33GVDP":3,"skill-same:01M71HVB9501BZKWRX5M33GVDP":73},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":38,"linkOnly":39,"markdownContent":40,"markdownBody":41,"frontMatter":42,"refName":46,"commitSha":47,"skillDirectory":48,"forksCount":49,"repositoryHtmlUrl":50,"repositoryHomepageUrl":6,"ownerHtmlUrl":51,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":52,"install":53,"resources":66},"01M71HVB9501BZKWRX5M33GVDP","minimax-music-playlist",null,"Generate personalized music playlists by analyzing the user's music taste and generation feedback history. Triggers on any request involving playlist generation, music taste profiling, or personalized music recommendations. Supports multilingual triggers — match equivalent phrases in any language.","xuanxuan1983\u002Fhermes-skills-backup","xuanxuan1983","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F151701653?v=4","01M71HCZNB5H19QJ6YBF98W2WH","01M71HCZP14T0921BD0TC9EPD3",1,"MIT","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fagent-skills\u002Fminimax-music-playlist\u002FSKILL.md",[],[18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37],"minimax music playlist","music","minimax","playlist","user","language","playlists","taste","match","analyzing","creative","request","recommendations","generate","profiling","generation","multilingual","triggers","history","feedback","2026-11-10 10:12:30",false,"---\nname: minimax-music-playlist\ndescription: >\n  Generate personalized music playlists by analyzing the user's music taste\n  and generation feedback history. Triggers\n  on any request involving playlist generation, music taste profiling, or personalized\n  music recommendations. Supports multilingual triggers — match equivalent phrases in\n  any language.\nlicense: MIT\nmetadata:\n  version: \"2.0\"\n  category: creative\n---\n\n# MiniMax Music Playlist — Personalized Playlist Generator\n\nScan the user's music taste, build a taste profile, generate a personalized\nplaylist, and create an album cover. This skill is designed for both agent and direct\nuser invocation — adapt interaction style to context.\n\n## Prerequisites\n\n- **mmx CLI** — music & image generation. Install: `npm install -g mmx-cli`. Auth: `mmx auth login --api-key \u003Ckey>`.\n- **Python 3** — for scanning scripts you write on the fly (stdlib only, no pip).\n- **Audio player** — `mpv`, `ffplay`, or `afplay` (macOS built-in).\n\n## Language\n\nDetect the user's language from their message. **All user-facing text must be\nin the same language as the user's prompt** — do not mix languages. If the user\nwrites in Chinese, all output (profile summary, theme suggestions, playlist plan,\nplayback info) must be fully in Chinese. If in English, all in English.\n\nAll `mmx` generation prompts should be in English for best quality.\nEach song's lyrics language follows its genre (K-pop → Korean, J-pop → Japanese, etc.),\nNOT the user's UI language.\n\n---\n\n## Workflow\n\n```\n1. Scan local music apps → 2. Build taste profile → 3. Plan playlist\n→ 4. Generate songs (mmx music) → 5. Generate cover (mmx image) → 6. Play → 7. Save & feedback\n```\n\n---\n\n## Step 1: Gather Music Listening Data\n\nCollect the user's listening data from available sources.\n\n**Supported sources:**\n\n| Source | Method | Data format |\n|--------|--------|-------------|\n| Apple Music | `osascript` to query Music.app (official AppleScript interface) | Track name, artist, album, genre, play count |\n| Spotify | User exports their own data via [Spotify Privacy Settings](https:\u002F\u002Fwww.spotify.com\u002Faccount\u002Fprivacy\u002F) | JSON files in ZIP (`Streaming_History_Audio_*.json`) |\n| Manual input | User describes their taste directly | Free text |\n\n**Spotify data export flow:**\nSpotify does not store useful data locally. To include Spotify listening history,\nfirst check if the user already has a Spotify data export:\n\n1. Search for existing exports: `find ~ -maxdepth 4 -name \"my_spotify_data.zip\" -o -name \"Streaming_History_Audio_*.json\" 2>\u002Fdev\u002Fnull`\n2. If found, ask the user if they want to use it\n3. If ZIP, unzip and locate `Spotify Extended Streaming History\u002FStreaming_History_Audio_*.json`\n4. If not found, open the Spotify privacy page: `open https:\u002F\u002Fwww.spotify.com\u002Faccount\u002Fprivacy\u002F`\n5. Tell the user to log in, scroll to \"Download your data\", and click \"Request data\"\n6. Skip Spotify for now and continue with other sources — tell the user they can\n   re-run the playlist skill after the data export arrives (usually a few days)\n\n**Spotify data format:**\nThe export contains `Streaming_History_Audio_YYYY.json` files (one per year), each\nis a JSON array of listening events. Key fields to extract:\n- `master_metadata_album_artist_name` — artist name\n- `master_metadata_track_name` — track name\n- `master_metadata_album_album_name` — album name\n- `ms_played` — playback duration in milliseconds (use as weight: longer = stronger signal)\n- `ts` — timestamp\n\nFilter out entries where `ms_played \u003C 30000` (less than 30 seconds, likely skipped).\nDo NOT use or store `ip_addr` or other sensitive fields.\n\n**What to extract from each source:**\n- Track names + artist names (primary signal)\n- Playlist names and membership (e.g., a playlist named \"Chinese Traditional\" tells you genre preference)\n- Play counts or streaming duration if available (weight frequently played tracks higher)\n- Scene\u002Fmood tags if available\n\n**Approach:**\n1. Check if Apple Music is available (try `osascript` query)\n2. Ask if the user has a Spotify data export ZIP to provide\n3. If no sources available, ask the user to describe their taste manually\n\n**Privacy rule:** Never show raw track lists to the user. Only show aggregated stats.\n\n---\n\n## Step 2: Build Taste Profile\n\nFrom the scanned data, build a taste profile covering:\n\n- **Genre distribution** — what styles the user listens to (e.g., J-pop 20%, R&B 15%, Classical 10%)\n- **Mood tendencies** — emotional tone preferences (melancholic, energetic, calm, romantic, etc.)\n- **Vocal preference** — male vs female voice ratio\n- **Tempo preference** — slow \u002F moderate \u002F upbeat \u002F fast distribution\n- **Language distribution** — zh, en, ja, ko, etc.\n- **Top artists** — most listened artists\n\n**How to infer genre\u002Fmood from artist names:**\nMost raw data only has artist + track names without genre tags. To enrich this:\n1. Look up artists in the local mapping table at `\u003CSKILL_DIR>\u002Fdata\u002Fartist_genre_map.json`\n   — this table covers 20,000 popular artists with pre-mapped genres, vocal type, and language\n2. For artists not in the mapping table, query the MusicBrainz API:\n   `https:\u002F\u002Fmusicbrainz.org\u002Fws\u002F2\u002Fartist\u002F?query=artist:\u003Cname>&fmt=json`\n   — extract genre tags from the response; respect rate limit (1 req\u002Fsec)\n   — cache results to `\u003CSKILL_DIR>\u002Fdata\u002Fartist_cache.json` to avoid re-querying\n3. If MusicBrainz returns no results, skip the artist\n\n**Profile caching:**\n- Save profile to `\u003CSKILL_DIR>\u002Fdata\u002Ftaste_profile.json`\n- If a profile less than 7 days old exists, reuse it (offer rescan option)\n- If older or missing, rebuild\n\n**Show user a summary:**\n```\nYour Music Profile:\n  Sources: Apple Music 230 | Spotify 140\n  Genres: J-pop 20% | R&B 15% | Classical 10% | Indie Pop 9%\n  Moods: Melancholic 25% | Calm 20% | Romantic 18%\n  Vocals: Female 65% | Male 35%\n  Top artists: Faye Wong, Ryuichi Sakamoto, Taylor Swift, Jay Chou, Taeko Onuki\n```\n\nIf invoked by an agent with clear parameters, skip the confirmation and proceed.\nIf invoked by a user directly, ask if the profile looks right before continuing.\n\n---\n\n## Step 3: Plan Playlist\n\n**Ask the user for a theme\u002Fscene before generating.** This is the one\ninteractive step in the workflow. All other steps run autonomously.\n\nIf the theme was already provided in the invocation (e.g., the agent or user\nsaid \"generate a late night chill playlist\"), use it directly and skip the question.\nOtherwise, ask:\n\n```\nWhat theme would you like for your playlist? Here are some suggestions:\n\n- \"Late night chill\" — relaxing slow songs\n- \"Commute\" — upbeat and energizing\n- \"Rainy day\" — melancholic & cozy\n- \"Surprise me\" — random based on your taste\n\nOr tell me your own vibe!\n```\n\nOnce the user picks a theme, proceed automatically through generation, cover,\nplayback, and saving — no further confirmations needed.\n\nDetermine playlist parameters:\n- **Theme\u002Fmood** — from user input, or default to top mood from profile\n- **Song count** — from user input, or default to 5\n- **Genre mix** — weighted by profile, with variety\n\n**Per-song lyrics language** follows genre:\n\n| Genre | Lyrics language |\n|-------|----------------|\n| K-pop, Korean R&B\u002Fballad | Korean |\n| J-pop, city pop, J-rock | Japanese |\n| C-pop, Chinese-style, Mandopop | Chinese |\n| Western pop\u002Findie\u002Frock\u002Fjazz\u002FR&B | English |\n| Latin pop, bossa nova | Spanish\u002FPortuguese |\n| Instrumental, lo-fi, ambient | No lyrics (`--instrumental`) |\n\nEmbed language naturally into the mmx prompt via vocal description:\n- Good: `\"A melancholy Chinese R&B ballad with a gentle introspective male voice, electric piano, bass, slow tempo\"`\n- Bad: `\"R&B ballad, melancholy... sung in Chinese\"`\n\n**Show the playlist plan before generating.** Display each song with two lines:\nthe first line shows genre, mood, and vocal\u002Flanguage tag; the second line shows\na short description of the song. **All user-facing text (plan, descriptions, moods,\nlabels) must be in the same language as the user's prompt.** Only the actual `--prompt`\npassed to `mmx` should be in English — this is internal and should NOT be shown to\nthe user. Example:\n\n```\nPlaylist Plan: Late Night Chill (5 songs)\n\n1. Neo-soul R&B — introspective  English\u002Fmale vocal\n   A mellow neo-soul R&B ballad with warm baritone, electric piano, smooth bass\n\n2. Lo-fi hip-hop — dreamy  Instrumental\n   Dreamy lo-fi with sampled piano, vinyl crackle, soft electronic drums\n\n3. Smooth jazz — romantic  English\u002Ffemale vocal\n   Silky female voice, saxophone, piano, romantic starlit night\n\n4. Indie folk — melancholic  English\u002Fmale vocal\n   Tender male voice, acoustic guitar, harmonica, quiet solitude\n\n5. Ambient electronic — calm  Instrumental\n   Soft synth pads, gentle arpeggios, dreamy atmosphere\n```\n\nAfter showing the plan, proceed directly to generation — no confirmation needed.\nThe user has already chosen the theme; the plan is shown for transparency, not approval.\n\n---\n\n## Step 4: Generate Songs\n\nUse `mmx music generate` to create all songs. **Generate concurrently** (up to 5 in parallel).\n\n```bash\n# Example: 5 songs in parallel\nmmx music generate --prompt \"\u003Cenglish_prompt_1>\" --lyrics-optimizer \\\n  --out ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cname>\u002F01_desc.mp3 --quiet --non-interactive &\nmmx music generate --prompt \"\u003Cenglish_prompt_2>\" --instrumental \\\n  --out ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cname>\u002F02_desc.mp3 --quiet --non-interactive &\n# ... more songs ...\nwait\n```\n\n**Key flags:**\n- `--lyrics-optimizer` — auto-generate lyrics from prompt (for vocal tracks)\n- `--instrumental` — no vocals\n- `--vocals \"\u003Cdescription>\"` — vocal style (e.g., \"warm Chinese male baritone\")\n- `--genre`, `--mood`, `--tempo`, `--instruments` — fine-grained control\n- `--quiet --non-interactive` — suppress interactive output for batch mode\n- `--out \u003Cpath>` — save to file\n\n**File naming:** `\u003CNN>_\u003Cshort_desc>.mp3` (e.g., `01_rnb_midnight.mp3`)\n\n**Output directory:** `~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cplaylist_name>\u002F`\n\nIf a song fails, **retry once** before skipping. Log the error and continue with the rest.\n\n---\n\n## Step 5: Generate Album Cover\n\nGenerate the album cover **concurrently with the songs** (Step 4), not after.\nLaunch the `mmx image generate` call in parallel with the song generation calls.\n\nCraft a prompt that reflects the playlist's theme, mood, and genre mix. The image\nshould feel like an album cover — artistic, evocative, not literal.\n\n```bash\nmmx image generate \\\n  --prompt \"\u003Ccover description based on playlist theme and mood>\" \\\n  --aspect-ratio 1:1 \\\n  --out-dir ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cplaylist_name>\u002F \\\n  --out-prefix cover \\\n  --quiet\n```\n\n**Prompt guidance:**\n- Abstract\u002Fartistic style works best for album covers\n- Reference the dominant mood and genre (e.g., \"dreamy late-night cityscape, neon reflections, lo-fi aesthetic\")\n- Do NOT include text or song titles in the image prompt\n- Aspect ratio should be 1:1 (square, standard album cover)\n\n---\n\n## Step 6: Playback\n\nDetect an available player and play the playlist in order:\n\n| Player | Command | Controls |\n|--------|---------|----------|\n| mpv | `mpv --no-video \u003Cfile>` | `q` skip, Space pause, arrows seek |\n| ffplay | `ffplay -nodisp -autoexit \u003Cfile>` | `q` skip |\n| afplay | `afplay \u003Cfile>` | Ctrl+C skip |\n\nPlay all `.mp3` files in the playlist directory in filename order.\nOnly play the songs generated in this session — if the directory has old files\nfrom a previous run, clean them out first or filter by the known filenames.\nIf no player is found, just show the file paths.\n\n---\n\n## Step 7: Save & Feedback\n\nSave playlist metadata to `\u003Cplaylist_dir>\u002Fplaylist.json`:\n```json\n{\n  \"name\": \"Late Night Chill\",\n  \"theme\": \"late night chill\",\n  \"created_at\": \"2026-04-11T22:00:00\",\n  \"song_count\": 5,\n  \"cover\": \"cover_001.png\",\n  \"songs\": [\n    {\"index\": 1, \"filename\": \"01_rnb_midnight.mp3\", \"prompt\": \"...\", \"rating\": null}\n  ]\n}\n```\n\nIf the user is present, ask for feedback (per-song or overall). Update the\ntaste profile's feedback section with liked\u002Fdisliked genres and prompts to\nimprove future playlists.\n\n---\n\n## Replaying Playlists\n\nIf asked to play a previous playlist: `ls ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F`, show\navailable ones, and play the selected one.\n\n---\n\n## Notes\n\n- **Agent vs user invocation**: The theme\u002Fscene question (Step 3) is the single\n  interactive touchpoint. If the theme is already provided in the invocation,\n  skip the question. Everything else runs autonomously.\n- **No hardcoded scripts**: Write scanning\u002Fanalysis scripts on the fly as needed.\n  Use Python stdlib only. Cache results to avoid redundant work.\n- **Skill directory**: `\u003CSKILL_DIR>` = the directory containing this SKILL.md file.\n  Data\u002Fcache files go in `\u003CSKILL_DIR>\u002Fdata\u002F`.\n- **All mmx prompts in English** for best generation quality.\n","# MiniMax Music Playlist — Personalized Playlist Generator\n\nScan the user's music taste, build a taste profile, generate a personalized\nplaylist, and create an album cover. This skill is designed for both agent and direct\nuser invocation — adapt interaction style to context.\n\n## Prerequisites\n\n- **mmx CLI** — music & image generation. Install: `npm install -g mmx-cli`. Auth: `mmx auth login --api-key \u003Ckey>`.\n- **Python 3** — for scanning scripts you write on the fly (stdlib only, no pip).\n- **Audio player** — `mpv`, `ffplay`, or `afplay` (macOS built-in).\n\n## Language\n\nDetect the user's language from their message. **All user-facing text must be\nin the same language as the user's prompt** — do not mix languages. If the user\nwrites in Chinese, all output (profile summary, theme suggestions, playlist plan,\nplayback info) must be fully in Chinese. If in English, all in English.\n\nAll `mmx` generation prompts should be in English for best quality.\nEach song's lyrics language follows its genre (K-pop → Korean, J-pop → Japanese, etc.),\nNOT the user's UI language.\n\n---\n\n## Workflow\n\n```\n1. Scan local music apps → 2. Build taste profile → 3. Plan playlist\n→ 4. Generate songs (mmx music) → 5. Generate cover (mmx image) → 6. Play → 7. Save & feedback\n```\n\n---\n\n## Step 1: Gather Music Listening Data\n\nCollect the user's listening data from available sources.\n\n**Supported sources:**\n\n| Source | Method | Data format |\n|--------|--------|-------------|\n| Apple Music | `osascript` to query Music.app (official AppleScript interface) | Track name, artist, album, genre, play count |\n| Spotify | User exports their own data via [Spotify Privacy Settings](https:\u002F\u002Fwww.spotify.com\u002Faccount\u002Fprivacy\u002F) | JSON files in ZIP (`Streaming_History_Audio_*.json`) |\n| Manual input | User describes their taste directly | Free text |\n\n**Spotify data export flow:**\nSpotify does not store useful data locally. To include Spotify listening history,\nfirst check if the user already has a Spotify data export:\n\n1. Search for existing exports: `find ~ -maxdepth 4 -name \"my_spotify_data.zip\" -o -name \"Streaming_History_Audio_*.json\" 2>\u002Fdev\u002Fnull`\n2. If found, ask the user if they want to use it\n3. If ZIP, unzip and locate `Spotify Extended Streaming History\u002FStreaming_History_Audio_*.json`\n4. If not found, open the Spotify privacy page: `open https:\u002F\u002Fwww.spotify.com\u002Faccount\u002Fprivacy\u002F`\n5. Tell the user to log in, scroll to \"Download your data\", and click \"Request data\"\n6. Skip Spotify for now and continue with other sources — tell the user they can\n   re-run the playlist skill after the data export arrives (usually a few days)\n\n**Spotify data format:**\nThe export contains `Streaming_History_Audio_YYYY.json` files (one per year), each\nis a JSON array of listening events. Key fields to extract:\n- `master_metadata_album_artist_name` — artist name\n- `master_metadata_track_name` — track name\n- `master_metadata_album_album_name` — album name\n- `ms_played` — playback duration in milliseconds (use as weight: longer = stronger signal)\n- `ts` — timestamp\n\nFilter out entries where `ms_played \u003C 30000` (less than 30 seconds, likely skipped).\nDo NOT use or store `ip_addr` or other sensitive fields.\n\n**What to extract from each source:**\n- Track names + artist names (primary signal)\n- Playlist names and membership (e.g., a playlist named \"Chinese Traditional\" tells you genre preference)\n- Play counts or streaming duration if available (weight frequently played tracks higher)\n- Scene\u002Fmood tags if available\n\n**Approach:**\n1. Check if Apple Music is available (try `osascript` query)\n2. Ask if the user has a Spotify data export ZIP to provide\n3. If no sources available, ask the user to describe their taste manually\n\n**Privacy rule:** Never show raw track lists to the user. Only show aggregated stats.\n\n---\n\n## Step 2: Build Taste Profile\n\nFrom the scanned data, build a taste profile covering:\n\n- **Genre distribution** — what styles the user listens to (e.g., J-pop 20%, R&B 15%, Classical 10%)\n- **Mood tendencies** — emotional tone preferences (melancholic, energetic, calm, romantic, etc.)\n- **Vocal preference** — male vs female voice ratio\n- **Tempo preference** — slow \u002F moderate \u002F upbeat \u002F fast distribution\n- **Language distribution** — zh, en, ja, ko, etc.\n- **Top artists** — most listened artists\n\n**How to infer genre\u002Fmood from artist names:**\nMost raw data only has artist + track names without genre tags. To enrich this:\n1. Look up artists in the local mapping table at `\u003CSKILL_DIR>\u002Fdata\u002Fartist_genre_map.json`\n   — this table covers 20,000 popular artists with pre-mapped genres, vocal type, and language\n2. For artists not in the mapping table, query the MusicBrainz API:\n   `https:\u002F\u002Fmusicbrainz.org\u002Fws\u002F2\u002Fartist\u002F?query=artist:\u003Cname>&fmt=json`\n   — extract genre tags from the response; respect rate limit (1 req\u002Fsec)\n   — cache results to `\u003CSKILL_DIR>\u002Fdata\u002Fartist_cache.json` to avoid re-querying\n3. If MusicBrainz returns no results, skip the artist\n\n**Profile caching:**\n- Save profile to `\u003CSKILL_DIR>\u002Fdata\u002Ftaste_profile.json`\n- If a profile less than 7 days old exists, reuse it (offer rescan option)\n- If older or missing, rebuild\n\n**Show user a summary:**\n```\nYour Music Profile:\n  Sources: Apple Music 230 | Spotify 140\n  Genres: J-pop 20% | R&B 15% | Classical 10% | Indie Pop 9%\n  Moods: Melancholic 25% | Calm 20% | Romantic 18%\n  Vocals: Female 65% | Male 35%\n  Top artists: Faye Wong, Ryuichi Sakamoto, Taylor Swift, Jay Chou, Taeko Onuki\n```\n\nIf invoked by an agent with clear parameters, skip the confirmation and proceed.\nIf invoked by a user directly, ask if the profile looks right before continuing.\n\n---\n\n## Step 3: Plan Playlist\n\n**Ask the user for a theme\u002Fscene before generating.** This is the one\ninteractive step in the workflow. All other steps run autonomously.\n\nIf the theme was already provided in the invocation (e.g., the agent or user\nsaid \"generate a late night chill playlist\"), use it directly and skip the question.\nOtherwise, ask:\n\n```\nWhat theme would you like for your playlist? Here are some suggestions:\n\n- \"Late night chill\" — relaxing slow songs\n- \"Commute\" — upbeat and energizing\n- \"Rainy day\" — melancholic & cozy\n- \"Surprise me\" — random based on your taste\n\nOr tell me your own vibe!\n```\n\nOnce the user picks a theme, proceed automatically through generation, cover,\nplayback, and saving — no further confirmations needed.\n\nDetermine playlist parameters:\n- **Theme\u002Fmood** — from user input, or default to top mood from profile\n- **Song count** — from user input, or default to 5\n- **Genre mix** — weighted by profile, with variety\n\n**Per-song lyrics language** follows genre:\n\n| Genre | Lyrics language |\n|-------|----------------|\n| K-pop, Korean R&B\u002Fballad | Korean |\n| J-pop, city pop, J-rock | Japanese |\n| C-pop, Chinese-style, Mandopop | Chinese |\n| Western pop\u002Findie\u002Frock\u002Fjazz\u002FR&B | English |\n| Latin pop, bossa nova | Spanish\u002FPortuguese |\n| Instrumental, lo-fi, ambient | No lyrics (`--instrumental`) |\n\nEmbed language naturally into the mmx prompt via vocal description:\n- Good: `\"A melancholy Chinese R&B ballad with a gentle introspective male voice, electric piano, bass, slow tempo\"`\n- Bad: `\"R&B ballad, melancholy... sung in Chinese\"`\n\n**Show the playlist plan before generating.** Display each song with two lines:\nthe first line shows genre, mood, and vocal\u002Flanguage tag; the second line shows\na short description of the song. **All user-facing text (plan, descriptions, moods,\nlabels) must be in the same language as the user's prompt.** Only the actual `--prompt`\npassed to `mmx` should be in English — this is internal and should NOT be shown to\nthe user. Example:\n\n```\nPlaylist Plan: Late Night Chill (5 songs)\n\n1. Neo-soul R&B — introspective  English\u002Fmale vocal\n   A mellow neo-soul R&B ballad with warm baritone, electric piano, smooth bass\n\n2. Lo-fi hip-hop — dreamy  Instrumental\n   Dreamy lo-fi with sampled piano, vinyl crackle, soft electronic drums\n\n3. Smooth jazz — romantic  English\u002Ffemale vocal\n   Silky female voice, saxophone, piano, romantic starlit night\n\n4. Indie folk — melancholic  English\u002Fmale vocal\n   Tender male voice, acoustic guitar, harmonica, quiet solitude\n\n5. Ambient electronic — calm  Instrumental\n   Soft synth pads, gentle arpeggios, dreamy atmosphere\n```\n\nAfter showing the plan, proceed directly to generation — no confirmation needed.\nThe user has already chosen the theme; the plan is shown for transparency, not approval.\n\n---\n\n## Step 4: Generate Songs\n\nUse `mmx music generate` to create all songs. **Generate concurrently** (up to 5 in parallel).\n\n```bash\n# Example: 5 songs in parallel\nmmx music generate --prompt \"\u003Cenglish_prompt_1>\" --lyrics-optimizer \\\n  --out ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cname>\u002F01_desc.mp3 --quiet --non-interactive &\nmmx music generate --prompt \"\u003Cenglish_prompt_2>\" --instrumental \\\n  --out ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cname>\u002F02_desc.mp3 --quiet --non-interactive &\n# ... more songs ...\nwait\n```\n\n**Key flags:**\n- `--lyrics-optimizer` — auto-generate lyrics from prompt (for vocal tracks)\n- `--instrumental` — no vocals\n- `--vocals \"\u003Cdescription>\"` — vocal style (e.g., \"warm Chinese male baritone\")\n- `--genre`, `--mood`, `--tempo`, `--instruments` — fine-grained control\n- `--quiet --non-interactive` — suppress interactive output for batch mode\n- `--out \u003Cpath>` — save to file\n\n**File naming:** `\u003CNN>_\u003Cshort_desc>.mp3` (e.g., `01_rnb_midnight.mp3`)\n\n**Output directory:** `~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cplaylist_name>\u002F`\n\nIf a song fails, **retry once** before skipping. Log the error and continue with the rest.\n\n---\n\n## Step 5: Generate Album Cover\n\nGenerate the album cover **concurrently with the songs** (Step 4), not after.\nLaunch the `mmx image generate` call in parallel with the song generation calls.\n\nCraft a prompt that reflects the playlist's theme, mood, and genre mix. The image\nshould feel like an album cover — artistic, evocative, not literal.\n\n```bash\nmmx image generate \\\n  --prompt \"\u003Ccover description based on playlist theme and mood>\" \\\n  --aspect-ratio 1:1 \\\n  --out-dir ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F\u003Cplaylist_name>\u002F \\\n  --out-prefix cover \\\n  --quiet\n```\n\n**Prompt guidance:**\n- Abstract\u002Fartistic style works best for album covers\n- Reference the dominant mood and genre (e.g., \"dreamy late-night cityscape, neon reflections, lo-fi aesthetic\")\n- Do NOT include text or song titles in the image prompt\n- Aspect ratio should be 1:1 (square, standard album cover)\n\n---\n\n## Step 6: Playback\n\nDetect an available player and play the playlist in order:\n\n| Player | Command | Controls |\n|--------|---------|----------|\n| mpv | `mpv --no-video \u003Cfile>` | `q` skip, Space pause, arrows seek |\n| ffplay | `ffplay -nodisp -autoexit \u003Cfile>` | `q` skip |\n| afplay | `afplay \u003Cfile>` | Ctrl+C skip |\n\nPlay all `.mp3` files in the playlist directory in filename order.\nOnly play the songs generated in this session — if the directory has old files\nfrom a previous run, clean them out first or filter by the known filenames.\nIf no player is found, just show the file paths.\n\n---\n\n## Step 7: Save & Feedback\n\nSave playlist metadata to `\u003Cplaylist_dir>\u002Fplaylist.json`:\n```json\n{\n  \"name\": \"Late Night Chill\",\n  \"theme\": \"late night chill\",\n  \"created_at\": \"2026-04-11T22:00:00\",\n  \"song_count\": 5,\n  \"cover\": \"cover_001.png\",\n  \"songs\": [\n    {\"index\": 1, \"filename\": \"01_rnb_midnight.mp3\", \"prompt\": \"...\", \"rating\": null}\n  ]\n}\n```\n\nIf the user is present, ask for feedback (per-song or overall). Update the\ntaste profile's feedback section with liked\u002Fdisliked genres and prompts to\nimprove future playlists.\n\n---\n\n## Replaying Playlists\n\nIf asked to play a previous playlist: `ls ~\u002FMusic\u002Fminimax-gen\u002Fplaylists\u002F`, show\navailable ones, and play the selected one.\n\n---\n\n## Notes\n\n- **Agent vs user invocation**: The theme\u002Fscene question (Step 3) is the single\n  interactive touchpoint. If the theme is already provided in the invocation,\n  skip the question. Everything else runs autonomously.\n- **No hardcoded scripts**: Write scanning\u002Fanalysis scripts on the fly as needed.\n  Use Python stdlib only. Cache results to avoid redundant work.\n- **Skill directory**: `\u003CSKILL_DIR>` = the directory containing this SKILL.md file.\n  Data\u002Fcache files go in `\u003CSKILL_DIR>\u002Fdata\u002F`.\n- **All mmx prompts in English** for best generation quality.\n",{"name":5,"license":14,"metadata":43,"description":45},{"version":44,"category":28},"2.0","Generate personalized music playlists by analyzing the user's music taste and generation feedback history. Triggers on any request involving playlist generation, music taste profiling, or personalized music recommendations. Supports multilingual triggers — match equivalent phrases in any language.\n","main","17ee50c38aefa9426a89d7e8ee167fdc4c184fb7","agent-skills\u002Fminimax-music-playlist",0,"https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983","2026-05-24 00:14:48",{"source":8,"skillName":5,"commands":54,"prompt":64,"zipAvailable":65},[55,58,61],{"tool":56,"command":57},"npx","npx skills add xuanxuan1983\u002Fhermes-skills-backup --skill minimax-music-playlist",{"tool":59,"command":60},"bunx","bunx skills add xuanxuan1983\u002Fhermes-skills-backup --skill minimax-music-playlist",{"tool":62,"command":63},"pnpm","pnpm dlx skills add xuanxuan1983\u002Fhermes-skills-backup --skill minimax-music-playlist","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fagent-skills\u002Fminimax-music-playlist\u002FSKILL.md",true,[67],{"path":68,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":71,"sourceHtmlUrl":72},"data\u002Fartist_genre_map.json","REFERENCE","参考文档",3187972,"https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fagent-skills\u002Fminimax-music-playlist\u002Fdata\u002Fartist_genre_map.json",[74,117,145,161,183,204],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":96,"publishedAt":38},"01M71JFHVW1W4XE013TY9ZYN9P","writing-plans","Use when you have a spec or requirements for a multi-step task. Creates comprehensive implementation plans with bite-sized tasks, exact file paths, and complete code examples.","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fskills\u002Fsoftware-development\u002Fwriting-plans\u002FSKILL.md",[80,85,88,93],{"categoryNo":81,"categoryName":82,"taxonomyType":83,"taxonomyTypeDesc":84},"01JBSEEDCAT000000000000015","视觉设计","SCENARIO","使用场景",{"categoryNo":86,"categoryName":87,"taxonomyType":83,"taxonomyTypeDesc":84},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":89,"categoryName":90,"taxonomyType":91,"taxonomyTypeDesc":92},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",{"categoryNo":94,"categoryName":95,"taxonomyType":91,"taxonomyTypeDesc":92},"01JBSEEDCAT000000000000036","内容创作者",[97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116],"writing","writing plans","plans","test-driven","test-driven-development","superpowers","design","when","workflow","you","code","tasks","subagent","code-review","review","documentation","paths","planning","development","for",{"skillNo":118,"name":101,"nameZh":6,"description":119,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":120,"categories":121,"keywords":130,"publishedAt":38},"01M71JFF859RV0XGWS206X1TMB","Use when implementing any feature or bugfix, before writing implementation code. Enforces RED-GREEN-REFACTOR cycle with test-first approach.","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fskills\u002Fsoftware-development\u002Ftest-driven-development\u002FSKILL.md",[122,125,128,129],{"categoryNo":123,"categoryName":124,"taxonomyType":83,"taxonomyTypeDesc":84},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":126,"categoryName":127,"taxonomyType":83,"taxonomyTypeDesc":84},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":86,"categoryName":87,"taxonomyType":83,"taxonomyTypeDesc":84},{"categoryNo":94,"categoryName":95,"taxonomyType":91,"taxonomyTypeDesc":92},[131,115,132,133,134,135,136,137,138,139,102,97,140,104,107,109,141,142,143,144],"test driven development","test","driven","test-first","green","cycle","red-green-refactor","testing","implementing","debugging","systematic","bugfix","refactor","systematic-debugging",{"skillNo":146,"name":144,"nameZh":6,"description":147,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":148,"categories":149,"keywords":151,"publishedAt":38},"01M71JFCA5S48PZ5M7DG31R8K5","Use when encountering any bug, test failure, or unexpected behavior. 4-phase root cause investigation — NO fixes without understanding the problem first.","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fskills\u002Fsoftware-development\u002Fsystematic-debugging\u002FSKILL.md",[150],{"categoryNo":126,"categoryName":127,"taxonomyType":83,"taxonomyTypeDesc":84},[152,140,141,100,153,154,101,102,97,104,109,155,156,115,157,158,99,159,160,132],"systematic debugging","problem","problem-solving","phase","bug","fixes","root cause","use","root-cause",{"skillNo":162,"name":163,"nameZh":6,"description":164,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":165,"categories":166,"keywords":176,"publishedAt":38},"01M71JF9RQ092KM5ZXCBBT6FS1","subagent-driven-development","Use when executing implementation plans with independent tasks. Dispatches fresh delegate_task per task with two-stage review (spec compliance then code quality).","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fskills\u002Fsoftware-development\u002Fsubagent-driven-development\u002FSKILL.md",[167,168,171,172,173],{"categoryNo":86,"categoryName":87,"taxonomyType":83,"taxonomyTypeDesc":84},{"categoryNo":169,"categoryName":170,"taxonomyType":83,"taxonomyTypeDesc":84},"01JBSEEDCAT000000000000023","Skill与智能体开发",{"categoryNo":89,"categoryName":90,"taxonomyType":91,"taxonomyTypeDesc":92},{"categoryNo":94,"categoryName":95,"taxonomyType":91,"taxonomyTypeDesc":92},{"categoryNo":174,"categoryName":175,"taxonomyType":91,"taxonomyTypeDesc":92},"01JBSEEDCAT000000000000037","律师与法务",[177,109,115,133,100,101,178,179,102,97,104,105,107,108,110,111,180,181,182,99],"subagent driven development","delegate","delegation","fresh","parallel","code quality",{"skillNo":184,"name":185,"nameZh":6,"description":186,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":187,"categories":188,"keywords":191,"publishedAt":38},"01M71JF76DA8ARZ4JRYR4VBRWM","requesting-code-review","Pre-commit verification pipeline — static security scan, baseline-aware quality gates, independent reviewer subagent, and auto-fix loop. Use after code changes and before committing, pushing, or opening a PR.","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fskills\u002Fsoftware-development\u002Frequesting-code-review\u002FSKILL.md",[189,190],{"categoryNo":123,"categoryName":124,"taxonomyType":83,"taxonomyTypeDesc":84},{"categoryNo":89,"categoryName":90,"taxonomyType":91,"taxonomyTypeDesc":92},[192,193,107,111,100,194,195,101,196,197,102,198,97,199,200,201,109,110,202,203],"code review","requesting code review","pipeline","static","fix","auto-fix","auto","loop","security scan","commit","reviewer","verification",{"skillNo":205,"name":206,"nameZh":6,"description":207,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":208,"categories":209,"keywords":210,"publishedAt":38},"01M71JF52HZT5VQXFVF9K2H76K","plan","Plan mode for Hermes — inspect context, write a markdown plan into the active workspace's `.hermes\u002Fplans\u002F` directory, and do not execute the work.","https:\u002F\u002Fgithub.com\u002Fxuanxuan1983\u002Fhermes-skills-backup\u002Fblob\u002Fmain\u002Fskills\u002Fsoftware-development\u002Fplan\u002FSKILL.md",[],[206,211,212,114,115,116,213,214,99,97,105,109,215,216,217,218,219,220,221,222],"plan-mode","do","inspect","mode","work","markdown","active","directory","workspace","plan mode","write","implementation"]