Openclaw-ecommerce-ad-skills story-video-orchestrator
端到端剧情视频流水线编排,串联角色图、主题、分镜生成与质量分析,适合漫剧、剧情短片和角色一致性视频创作。
install
source · Clone the upstream repo
git clone https://github.com/lujiaheng-artpivot/openclaw-ecommerce-ad-skills
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/lujiaheng-artpivot/openclaw-ecommerce-ad-skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/story-video-orchestrator" ~/.claude/skills/lujiaheng-artpivot-openclaw-ecommerce-ad-skills-story-video-orchestrator && rm -rf "$T"
OpenClaw · Install into ~/.openclaw/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/lujiaheng-artpivot/openclaw-ecommerce-ad-skills "$T" && mkdir -p ~/.openclaw/skills && cp -r "$T/skills/story-video-orchestrator" ~/.openclaw/skills/lujiaheng-artpivot-openclaw-ecommerce-ad-skills-story-video-orchestrator && rm -rf "$T"
manifest:
skills/story-video-orchestrator/SKILL.mdsource content
剧情视频流水线 Skill
把
manga-drama 和 volcengine-video-understanding 组合成可复用的剧情视频流水线:角色图 -> 剧本/分镜 -> 多场景视频 -> 批量分析。
何时使用
- 用户要从一张角色图快速生成剧情短片
- 需要把主题、风格、分镜数量封装成一次执行
- 需要在生成后自动拿到漫剧分析报告
使用方法
python3 scripts/run_story_pipeline.py \ --character ./character.png \ --theme "樱花树下的邂逅" \ --style anime_2d \ --scenes 4 \ --ratio 16:9 \ --output-dir ./story_output
参数说明
:角色图片路径,必填--character
:故事主题;与--theme
二选一--script-file
:自定义剧本路径--script-file
:漫画风格,默认--styleanime_2d
:分镜数,默认 4--scenes
:输出比例,默认--ratio16:9
:单镜头时长,默认 5--duration
:跳过分析报告--skip-analysis
:输出目录,默认--output-dir./story_output
输出结构
story_output/ ├── task_info.json ├── drama/ │ ├── script.txt │ ├── scene_1.mp4 │ └── drama_info.json ├── analysis/ │ └── drama_analysis.md └── events/ ├── event_story_pipeline_started_*.json ├── event_drama_generated_*.json └── event_video_analysis_completed_*.json
事件流
story.pipeline.starteddrama.generated
(如果未跳过分析)video.analysis.completed
进阶资料
- 需要选择不同流程模式时,读取
references/workflow-modes.md - 需要拆解剧情节奏时,进一步读取
../manga-drama/references/story-beats.md