AgentSkillsCN

music-recommender

当用户希望根据YouTube Music播放列表获取音乐推荐、分析自己的音乐品味、询问相似艺人或歌曲,或寻找符合自己风格的新音乐时,可使用此功能。

SKILL.md
--- frontmatter
name: music-recommender
description: Use when user wants music recommendations based on their YouTube Music playlists, wants to analyze their music taste, asks about similar artists or songs, or wants to discover new music matching their style

Music Playlist Analyzer & Recommender

Analyze YouTube Music playlists and provide personalized music recommendations.

Available Tools

Use these MCP tools from the youtube-music server:

ToolPurpose
youtube-music:get_my_playlistsList user's playlists
youtube-music:get_playlist_itemsGet songs in a playlist
youtube-music:analyze_playlistGet taste analysis (top artists, patterns)
youtube-music:search_musicSearch for music on YouTube

Process

1. Fetch Playlists

code
Use youtube-music:get_my_playlists to show available playlists

2. Analyze Selected Playlist

code
Use youtube-music:analyze_playlist with the playlist_id

3. Identify Patterns

From the analysis, identify:

  • Dominant artists: Who appears most frequently
  • Music style: Infer genres from artist names
  • Era/decade: Modern, classic, mixed
  • Energy level: Based on artist styles
  • Language/region: If applicable

4. Generate Recommendations

Use your music knowledge to recommend:

  • Similar artists the user might not know
  • Specific songs that match the playlist vibe
  • Adjacent genres to explore

For each recommendation explain why it matches their taste.

5. Optionally Search

code
Use youtube-music:search_music to find specific tracks and provide links

Output Format

Structure recommendations as:

code
## Your Music Profile
- Primary genres: [identified genres]
- Top artists: [from analysis]
- Vibe: [energy/mood description]

## Recommended Artists
1. **[Artist Name]** - [Why they match]
   - Try: "[Song 1]", "[Song 2]"

2. **[Artist Name]** - [Why they match]
   - Try: "[Song 1]", "[Song 2]"

## Discovery Suggestions
- [Genre/style] exploration: [specific recommendations]
- Deep cuts from artists you like: [specific songs]

Tips

  • Ask which playlist to analyze if user has multiple
  • Consider playlist name for context (workout, chill, party, etc.)
  • Mix well-known and lesser-known recommendations
  • Include YouTube Music search links when helpful