Video Script Generator

生成短视频脚本和分镜。输入产品/主题描述、目标受众、时长，输出完整脚本+分镜结构（JSON/YAML）。支持多种模板：痛点-解决、反转剧情、before-after。触发词：视频脚本、短视频文案、分镜、storyboard、video script。

Security Scan

The skill is an instruction-only short-video script/storyboard generator and its files, requirements, and instructions are coherent with that purpose.

Purpose & Capability

Name/description (video script + storyboard) match the included SKILL.md and template files; no unrelated binaries, credentials, or config paths are requested.

Instruction Scope

Runtime instructions only collect user inputs (topic, audience, duration, template, etc.) and produce structured YAML/JSON; they do not instruct the agent to read unrelated files, access environment variables, or transmit data to external endpoints.

Install Mechanism

No install spec and no code files to execute — this is instruction-only, so nothing is downloaded or written to disk during install.

Credentials

The skill declares no required environment variables or credentials; templates and examples do not request secrets or unrelated service tokens.

Persistence & Privilege

always:false and default autonomous invocation are normal. The skill does not request persistent system presence or modify other skills' configuration.

Assessment

This skill is instruction-only and appears to do exactly what it says: generate short-video scripts and shot-level storyboards. There are no requested credentials or installs, which reduces risk. Before using, avoid pasting any sensitive or proprietary secrets (API keys, private images, or internal URLs) into the input fields because generated scripts may include or echo that content. Also note the SKILL.md mentions downstream skills (digital-avatar, scene-video-generator, video-stitcher) — if you connect outputs to other skills, review those skills' permissions and credentials separately. Overall this skill is coherent and low-risk, but exercise normal caution about the content you provide to it or to any downstream skill.

短视频脚本+分镜生成

工作流程

1. 收集输入信息
2. 选择模板风格
3. 生成脚本+分镜结构
4. 输出结构化数据供下游消费

输入参数

参数	必填	说明
topic	✓	产品/主题描述
audience	✓	目标受众
duration	✓	时长：15s / 30s / 60s / 90s
template	-	模板：pain-solution / before-after / plot-twist
platform	-	平台：douyin / xiaohongshu / youtube-shorts
tone	-	语气：professional / casual / humorous / dramatic

模板说明

pain-solution（默认）

痛点引入 → 解决方案 → 效果展示 → CTA

before-after

对比前的状态 → 转折点 → 对比后的状态 → CTA

plot-twist

正常叙事 → 意外反转 → 揭示真相 → CTA

详见 templates/ 目录。

输出格式

meta:
  title: "脚本标题"
  duration: 30s
  platform: douyin
  template: pain-solution

script:
  hook: "开场钩子文案"
  
  scenes:
    - id: 1
      type: hook           # hook/pain/solution/demo/cta
      duration: 3s
      narration: "旁白/台词"
      shot_description: "镜头描述：特写/中景/全景，动作，表情"
      visual_style: "画面风格提示"
      demo_insert: false
      
    - id: 2
      type: pain
      duration: 5s
      narration: "..."
      shot_description: "..."
      demo_insert: false
      
    - id: 3
      type: solution
      duration: 8s
      narration: "..."
      shot_description: "产品界面展示"
      demo_insert: true
      demo_action: "点击生成按钮，展示结果"
      
    - id: 4
      type: cta
      duration: 3s
      narration: "链接在评论区"
      shot_description: "logo + 二维码"
      demo_insert: false

  cta: "链接在评论区 / 点击头像私信"
  
music_suggestion: "轻快节奏 / 紧张悬疑 / 温馨治愈"

场景类型

type	用途	典型时长
hook	开场吸引	2-3s
pain	痛点展示	3-5s
solution	解决方案	5-10s
demo	产品演示	5-15s
result	效果展示	3-5s
twist	反转	3-5s
cta	行动号召	2-3s

时长分配参考

总时长	场景数	分配建议
15s	3-4	hook(2) + core(10) + cta(3)
30s	4-6	hook(3) + pain(5) + solution(15) + cta(3)
60s	6-8	hook(3) + pain(8) + solution(35) + result(10) + cta(4)

使用示例

用户：帮我写一个30秒的短视频脚本，推广我的AI写作工具，目标用户是自媒体博主

执行：

1. 确认参数：topic=AI写作工具, audience=自媒体博主, duration=30s
2. 选择模板：pain-solution（默认）
3. 生成脚本+分镜
4. 输出 YAML 结构

下游对接

输出的结构化数据可直接用于：

• digital-avatar skill：提取 narration 生成口播
• scene-video-generator skill：按 shot_description 生成画面
• video-stitcher skill：按 duration 拼接