Agent Skills0xsline/OpenChatCut › music-intelligence

music-intelligence

GitHub

提供本地音乐分析、节拍检测及视频卡点剪辑与照片排版功能。支持通过计划工具生成编辑方案,并执行同步操作,适用于BGM节奏编辑和多媒体时间轴同步场景。

src/agent/skills/music-intelligence/SKILL.md 0xsline/OpenChatCut

Trigger Scenarios

用户请求分析音乐结构或节拍 需要进行视频卡点剪辑或按节拍放置照片

Install

npx skills add 0xsline/OpenChatCut --skill music-intelligence -g -y
More Options

Non-standard path

npx skills add https://github.com/0xsline/OpenChatCut/tree/main/src/agent/skills/music-intelligence -g -y

Use without installing

npx skills use 0xsline/OpenChatCut@music-intelligence

指定 Agent (Claude Code)

npx skills add 0xsline/OpenChatCut --skill music-intelligence -a claude-code -g -y

安装 repo 全部 skill

npx skills add 0xsline/OpenChatCut --all -g -y

预览 repo 内 skill

npx skills add 0xsline/OpenChatCut --list

SKILL.md

Frontmatter
{
    "name": "music-intelligence",
    "description": "Run or inspect local music analysis and plan or apply beat-, downbeat-, or section-synced video cuts or photo placements.\nUse for BGM beat edits, rhythm cuts, 卡点剪辑, musical structure, BPM, mood, genre, or instrument questions.\n",
    "user-invocable": false
}

Music Intelligence

  1. Call analyze_music when the user asks to analyze music or no cache exists. It runs the installed local Beat This + CLAP models, waits for completion, and never downloads models; use force:true only for an explicit reanalysis request.
  2. Call inspect_music when only the existing cache is needed.
  3. Before cutting video, call music_edit_plan and show its bounded cut/target summary. Before placing photos, call music_image_plan and show its bounded placement summary. Prefer timing:auto; choose sparse, medium, or dense from the requested pace.
  4. Only after the plan is accepted, call sync_cuts_to_music or sync_images_to_music with the returned analysisRef. Each tool recomputes the plan, rejects stale analysis, and applies its changes as one undo step.
  5. If required model packs are missing, report the returned install guidance, then retry analyze_music after installation.
  6. Never request, return, quote, summarize, or place the CLAP embedding in model context. Use tags, sections, confidence, and the opaque analysisRef only.

Preconditions

  • Analysis is opt-in and on-device. The media-pool asset must already have a cache for its current sourceRevision.
  • rhythm-lite supplies Beat This rhythm data; music-semantics-lite supplies CLAP tags and the private similarity vector.
  • An Agent call must never install a pack, download a model, decode media, or start inference. Direct the user to Settings and the media-card analysis action when the cache is unavailable.

Inspect

  • Identify the BGM by itemId, assetId, or an unambiguous name. Prefer the timeline item when the user intends to edit, because returned points are then mapped to timeline frames through the clip trim and playback rate.
  • inspect_music returns BPM, meter, confidence, tags, sections, counts, and bounded point lists. Lists are capped at 48 beats, 24 downbeats, 16 sections, and 12 tags; use a narrower range when truncated is true.
  • Treat analysisRef as opaque. It binds later planning and execution to the exact cached analysis without exposing the 512-value CLAP embedding.

Plan

  • timing:auto chooses section → downbeat → beat for sparse edits, downbeat → beat → section for medium edits, and beat → downbeat → section for dense edits.
  • Density samples every fourth, second, or first candidate for sparse, medium, or dense respectively.
  • Plans are bounded to 96 cut frames and 64 overlapping video targets. Report the returned summary and any cap or locked-target warning before applying.
  • A plan targets video clips that overlap the selected BGM range. It must never include or mutate the BGM item itself.

Photo placement

  • music_image_plan uses explicitly listed image asset ids in their given order, or all image assets in media-pool order when omitted. It cycles that order when the music range has more beat intervals than images.
  • Each placement covers the complete interval from one selected timing point to the next, including the BGM range boundaries, so images are not appended with their source duration or placed at a uniform default length.
  • Image placement is bounded to 96 clips and 64 image assets. sync_images_to_music requires an unlocked, empty target video track for the requested range; choose another track instead of overwriting existing content.

Apply and recover

  • Pass the same itemId, timing, density, range, target ids, and analysisRef to sync_cuts_to_music. Execution recomputes the plan from current state; do not replay an old list of frames.
  • A missing or stale ref means the project or analysis changed. Call inspect_music, then music_edit_plan or music_image_plan, and ask for acceptance again.
  • Locked video tracks are skipped and reported. Never unlock a track implicitly.
  • Successful splits are submitted as one EditorCommands.batch, so the whole rhythm edit is one undo step.
  • Successful photo placements are submitted as one EditorCommands.batch, so the whole sequence is one undo step.

Version History

  • d4052ab Current 2026-08-16 15:57

    新增基于节拍的照片排版工作流(music_image_plan/sync_images_to_music),完善缓存缺失时的错误指引,并扩展音乐分析元数据暴露。

  • 0c0c02d 2026-08-12 09:07

Same Skill Collection

src/agent/skills/ai-cinematic-short-film/SKILL.md
src/agent/skills/asset-import/SKILL.md
src/agent/skills/create-motion-graphics/SKILL.md
src/agent/skills/explainer-video/SKILL.md
src/agent/skills/export/SKILL.md
src/agent/skills/image-gen/SKILL.md
src/agent/skills/known-errors/SKILL.md
src/agent/skills/long-video-to-shorts/SKILL.md
src/agent/skills/motion-graphic-placement/SKILL.md
src/agent/skills/multi-clips-to-reels/SKILL.md
src/agent/skills/music/SKILL.md
src/agent/skills/news-rough-cut/SKILL.md
src/agent/skills/openchatcut-plugin-basics/SKILL.md
src/agent/skills/product-ad-video-script/SKILL.md
src/agent/skills/shader-gen/SKILL.md
src/agent/skills/storyboard-shot-breakdown/SKILL.md
src/agent/skills/transcription/SKILL.md
src/agent/skills/verification/SKILL.md
src/agent/skills/video-gen/SKILL.md
src/agent/skills/video-thumbnail-generator/SKILL.md
src/agent/skills/voice/SKILL.md
src/agent/skills/widget-forms/SKILL.md
skills/openchatcut/SKILL.md
src/agent/skills/product-help/SKILL.md
src/agent/skills/talking-head-guide/SKILL.md

Metadata

Files
0
Version
d4052ab
Hash
1a84851e
Indexed
2026-08-12 09:07

inicio - Wiki
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-08-16 19:39
浙ICP备14020137号-1 $mapa de visitantes$