Token导航 LogoToken导航

Agent Skill · AI 工具

video-designer(视频设计师)

用于辅助视频生成、动画合成、脚本化剪辑或 Remotion 等视频项目开发。它适合让 Agent 组织镜头、生成素材说明、维护合成代码或排查渲染问题。使用时需要确认分辨率、时长、素材路径和导出格式;涉及外部素材、人物肖像或商业发布时,应先核对版权授权和内容审核要求。

video-designer 是AI 工具类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证为 MIT,第三方扫描有提醒项,按安装量在AI 工具类的 2,536 个 Skill 里排第 1,931。

总安装 399周安装 16GitHub Stars 53下载 129核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:video-designer(视频设计师)
来源仓库:https://github.com/outscal/video-generator
仓库路径:skills/video-designer
安装命令:
npx skills add https://github.com/outscal/video-generator --skill video-designer
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/outscal/video-generator --skill video-designer

这个 Skill 是做什么的

AI 工具类 Skill 用来扩展助手本身的能力,例如接模型、接检索、接外部服务。

用于辅助视频生成、动画合成、脚本化剪辑或 Remotion 等视频项目开发。适合让 Agent 组织镜头、生成素材说明、维护合成代码或排查渲染问题。使用时需要确认分辨率、时长、素材路径和导出格式。涉及外部素材、人物肖像或商业发布时,应先核对版权授权和内容审核要求。

video-designer 属于AI 工具类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/video-designer

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/outscal/video-generator --skill video-designer;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

video-designer 怎么安装?

在终端执行 npx skills add https://github.com/outscal/video-generator --skill video-designer。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

video-designer 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

video-designer 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

video-designer 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

video-designer 用的人多吗?

公开数据是累计安装 399 次,近一周安装 16 次,下载 129 次,GitHub 53 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

video-designer 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理AI 工具类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

AI 工具类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
executing-plans

执行计划 · Codex、Claude

1,515,336173,110只读MIT通过
verification-before-completion

完成前的验证 · Codex、Claude

1,319,472173,121权限未标注MIT通过
subagent-driven-development

子 Agent 驱动开发 · Codex、Claude

1,312,416173,138只读MIT通过
copy-editing

复制编辑 · Codex、Claude

1,209,57625,632只读MIT通过
paywall-upgrade-cro

付费专区升级 cro · Codex、Claude

1,030,17625,617只读MIT通过
video-designer

视频设计师 · 本页

39953只读MIT有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Video Designer

./references/characters.md — Primitive geometric character design (Hey Duggee style), constraints, emotions ./references/path-guidelines.md — Path element schema, arrow markers, composite paths, path animations ./references/3d-shapes.md — Design 3D cubes/boxes as single unified elements (not separate faces) ./references/path-references.md — Different types of paths along with their parameters and output JSON example

COORDINATE SYSTEM
Origin: Top-left corner (0, 0)
X-axis: Increases rightward →
Y-axis: Increases downward ↓

FOR SHAPES/TEXT/ICONS:
  Position: Always refers to element's CENTER point

FOR PATHS:
  All coordinates are ABSOLUTE screen positions
  No position/size fields needed (implied by path coordinates)

ROTATION DIRECTIONS
0° = pointing up (↑)
90° = pointing right (→)
180° = pointing down (↓)
270° = pointing left (←)

Positive values = clockwise rotation
Negative values = counter-clockwise (-90° same as 270°)

SETTING TRANSFORMS BY ELEMENT TYPE

For shapes/text: Set `rotation` directly to the desired direction.

For assets: Check `asset_manifest` for `composition`.
- `composition` describes the asset's structure and orientation
- Infer the asset's base orientation from the composition description
- Use composition to determine if the asset is symmetric or asymmetric

SYMMETRIC ASSETS (no distinct top/bottom in composition):
- Use `rotation = desired_direction - inferred_orientation`

ASYMMETRIC ASSETS (composition mentions distinct top/bottom):
- If NO direction change needed, omit rotation/flip fields entirely
- If direction change is small (e.g., 45°) and won't invert, use `rotation` only
- If direction change would invert the asset (180°), use `flipX`/`flipY` instead of rotation
- `flipX: true` = horizontal mirror (for LEFT↔RIGHT flip)
- `flipY: true` = vertical mirror (for UP↔DOWN flip)
- Can combine flip + rotation for diagonal directions

Example: Asset with composition: "Wedge-shaped vehicle pointing RIGHT. Flat BOTTOM, angled TOP"
- Inferred orientation: 90° (pointing RIGHT)
- To point RIGHT: No transform needed (omit rotation/flip fields)
- To point UP-RIGHT: Use `rotation: -45` (small angle, no inversion)
- To point BOTTOM-RIGHT: Use `rotation: 45` (small angle, no inversion)
- To point LEFT: Use `flipX: true` (not rotation: 180 which inverts it)
- To point UP-LEFT: Use `flipX: true, rotation: 45`
- To point DOWN-LEFT: Use `flipX: true, rotation: -45`

EXAMPLE (1920×1080 viewport)
Screen center:        x = 960,  y = 540
Top-center:           x = 960,  y = 100
Bottom-left quadrant: x = 480,  y = 810
Right edge center:    x = 1820, y = 540

"video_metadata": {"viewport_size": "1920x1080", "backgroundColor": "#HEXCOLOR", "layout": {"strategy": "three-column|centered|freeform"}}, "elements": [{"id": "unique_element_id", "type": "shape|text|asset|pattern|path", "enterOn": 0, "exitOn": 7664, "content": "CRITICAL: Complete visual specification (see content_field_requirements)", "x": number, "y": number, "width": number, "height": number, "rotation": number, "flipX": boolean, "flipY": boolean, "orientation": number, "scale": number, "opacity": number, "fill": "#HEXCOLOR", "stroke": "#HEXCOLOR", "strokeWidth": number, "fontSize": number, "fontWeight": number, "textAlign": "left|center|right", "lineHeight": number, "zIndex": number, "animation": {"entrance": {"type": "string (e.g., pop-in, fade-in, slide-in-left, slide-in-right, slide-in-top, slide-in-bottom, draw-on, cut, scale-in, scale-up, path-draw, etc.)", "duration": number}, "exit": {"type": "string (e.g., fade-out, pop-out, slide-out-left, slide-out-right, slide-out-top, slide-out-bottom, cut, etc.)", "duration": number}, "actions": [{"on": number, "duration": number, "targetProperty": "opacity|scale|x|y|rotation|fill|stroke", "value": number, "easing": "linear|easeIn|easeOut|easeInOut"}, {"type": "follow-path", "pathId": "path_element_id", "autoRotate": boolean, "on": number, "duration": number, "easing": "linear|easeIn|easeOut|easeInOut"}]}}]}

**Required fields per element:** `id`, `type`, `enterOn`, `exitOn`, `content`, `zIndex`
**Required for assets with follow-path:** `orientation` (inferred from composition, e.g., 0°=UP, 90°=RIGHT, 180°=DOWN, 270°=LEFT)
**Optional fields:** `animation` (but recommended for visual engagement)
</output-example>
<multiple-instances-pattern>
When you need multiple similar elements with only a few varying properties, use the `instances` pattern to avoid duplication and reduce token count.
<syntax>
All base properties go at the root level. The `instances` array specifies overrides for each copy.

{ "id": "arrow", "type": "shape", "content": "Right-pointing arrow", "enterOn": 1000, "exitOn": 5000, "x": 100, "y": 200, "width": 50, "height": 50, "fill": "#3B82F6", "opacity": 1, "instances": [ { "useDefaults": true }, { "x": 200 }, { "x": 300, "fill": "#FF5722" }, { "y": 400 } ] }


**This creates 4 elements:**

- `arrow_0`: x=100, y=200, fill=#3B82F6 (uses all base values)
- `arrow_1`: x=200, y=200, fill=#3B82F6 (overrides only x)
- `arrow_2`: x=300, y=200, fill=#FF5722 (overrides x and fill)
- `arrow_3`: x=100, y=400, fill=#3B82F6 (overrides only y)

✅ **MUST use instances for:**

- Any 2+ elements of the same `type` with similar structure
- This includes: shapes, paths, text labels, assets - ANY element type

Creates 3 dots:

- `dot_0`: x=660, animation.entrance.duration=300 (base values)
- `dot_1`: x=960, animation.entrance.duration=200, type="pop-in" inherited (deep merge)
- `dot_2`: x=1260, fill=#FF5722, animation.entrance.duration=300 inherited

Creates 4 arc paths with different radii, positions, and directions while sharing the base coordinates.

- **Logical organization** - Related elements stay together (e.g., all parts of a smartphone)
- **Simplified management** - Transform entire groups at once (move, rotate, scale)
- **Shared timing** - Parent `enterOn`/`exitOn` applies to all children unless overridden
- **Reduced repetition** - Common properties defined at parent level

❌ **Don't use groups when:**

- Single standalone element
- Elements are unrelated or independent
- No shared timing or transformation needed

Individual children can override timing and have their own animations. **CRITICAL: Timing Values Must Be Absolute**

**All timing values (`enterOn`, `exitOn`, `action.on`) must use absolute video timestamps, NOT relative scene timestamps.**

Given:

- `scene.startTime = 18192` (absolute video time)
- Audio transcript shows word "dust" at `1777ms` (relative to scene start)

Your timing should be:

"enterOn": 19969, // 18192 + 1777 = absolute video time "exitOn": 24589 // matches scene.endTime (absolute)


**Formula:** `absolute_time = scene.startTime + audio_relative_time` The `content` field is the most critical part. It must answer ALL of these:

| Aspect | What to Specify | Example |
| --- | --- | --- |
| **Shape/Form** | Exact geometry, proportions | "Asymmetrical rounded blob—right lobe shorter, left lobe extends 2x downward" |
| **Visual Details** | Colors, textures, features | "Deep orange center (#E65100) fading to bright orange (#FF9800) edges, 3 subtle lighter spots" |
| **Face/Expression** | If character: eyes, mouth, emotion | "Wide white eyes with violet pupils, V-shaped pink eyebrows angled inward expressing anger" |
| **Position Context** | Where in frame, relative to what | "Centered in belly area of silhouette, taking 75% of belly's width" |
| **Initial State** | Starting appearance | "Begins as small concentrated core at liver's center" |
| **Transformations** | What changes and how | "On inhale: body compresses, eyes shrink, mouth tightens to small 'o'; on exhale: expands, eyes widen, mouth stretches to tall oval" |
| **Interaction** | How it relates to other elements | "Scales at same rate as silhouette to maintain relative position inside belly" |

**Precision Test**: Could someone draw this without seeing the original? If uncertain, add more detail. Before designing anything, analyze the example to establish your style guide:

- Background color
- Layout strategy

Audio: "bat" at 5908ms (relative) Element enterOn: 7000 + 5800 = 12800ms (absolute, with anticipation)

Always add scene.startTime to audio timestamps. </sync-to-transcript>

</phase-2-design-new-scene>

</design-process>

</designer-general-guidelines>

<text-guidelines>

<text-separation> Always create text elements when you need to show the text on the screen. </text-separation>

<auto-sizing> Text elements are auto-sized based on content and fontSize. Position elements with adequate spacing to avoid overlaps. </auto-sizing>

<text-align-clarification> CRITICAL: textAlign is CSS only, NOT positioning

The textAlign field controls how text lines flow within the text container. It does NOT change where the container is positioned.

Universal coordinate system rule applies:

  • x,y is ALWAYS the center point of the text element
  • textAlign affects internal text alignment, not container position

When textAlign matters:

  • Multi-line text (e.g., "MONKEY\nTALKS") - shorter lines align left/center/right within the container width
  • Single-line text with containerWidth set - text aligns within that fixed width

When textAlign has NO visible effect:

  • Single-line text without containerWidth - the container shrinks to fit the text exactly, leaving no extra space for alignment to matter

Example - Multi-line text at x=540 (viewport center):


textAlign: "center" textAlign: "left" ┌──────┐ ┌──────┐ │MONKEY│ │MONKEY│ │ TALKS│ │TALKS │ └──────┘ └──────┘ ↑ ↑ x=540 x=540

Both containers are centered on x=540. PHASED (longest line) fills container width. Only ARRAY (shorter line) shifts based on textAlign. </text-align-clarification>

<text-sizing-guidelines>

<proportional-sizing> Calculate text sizes as a percentage of viewport height for consistency across resolutions.

Calculation formula:


fontSize = viewport_height × percentage

Principle: Choose percentage based on:

  • Visual hierarchy (headlines larger than body text)
  • Readability requirements (ensure text is readable at target viewport size)
  • Direction requirements (follow what the direction specifies)
  • Context (technical content may need different sizing than casual content)

</proportional-sizing>

<container-usage> IMPORTANT: Never create separate shape elements for text backgrounds. Use the container object instead.

The container object gives text its own background, border, and padding. The background automatically sizes to fit the text content. </container-usage>

</text-sizing-guidelines>

<readability-rules>

  • Keep text within viewport boundaries

</readability-rules>

<text-schema>

{
    "id": "Unique identifier for this text element",
    "type": "Element type must be text",
    "content": "Brief description of what this text represents or its purpose in the scene",
    "text": "The actual text content to display",
    "bgID": "ID of parent/background element - ONLY for positioning text ON diagram elements (optional)",
    "enterOn": "ABSOLUTE video time in ms (scene.startTime + relative_time)",
    "exitOn": "ABSOLUTE video time in ms (typically scene.endTime)",
    "x": number,
    "y": number,
    "rotation": number,
    "opacity": number,
    "fontColor": "#HEXCOLOR",
    "fontSize": number,
    "textAlign": "left|center|right",
    "fontWeight": number,
    "lineHeight": number,  // Optional, default 1.5
    "containerWidth": number,  // Optional, for fixed-width text that wraps
    "padding": number,  // Optional, default 8
    "zIndex": number,

    "animation": {
        "entrance": {
        "type": "Entry animation style (string - e.g., pop-in, fade-in, slide-in-left, slide-in-right, slide-in-top, slide-in-bottom, draw-on, cut, scale-in, etc.)",
        "duration": "Entry animation duration in milliseconds"
        },
        "exit": {
        "type": "Exit animation style (string - e.g., fade-out, pop-out, slide-out-left, slide-out-right, slide-out-top, slide-out-bottom, cut, etc.)",
        "duration": "Exit animation duration in milliseconds"
        }
    },

       "container": {
        "padding": number,
        "background": {
        "type": "none|solid|gradient|frosted-glass|highlight",
           "color": "#HEXCOLOR",
           "opacity": number,
           "gradient": {
             "from": "#HEXCOLOR",
             "to": "#HEXCOLOR",
             "direction": "to-right|to-left|to-bottom|to-top|to-br|to-bl"
           }
         },
         "border": {
           "radius": number,
           "color": "#HEXCOLOR",
           "width": number
         },
         "backdropBlur": "sm|md|lg|xl"
       }
     }
{
  "id": "code_example_label",
  "type": "text",
  "content": "Label displaying code example with gradient background",
  "text": "Hello World!",
  "bgID": "",
  "enterOn": 1000,
  "exitOn": 5000,
  "x": 960,
  "y": 540,
  "rotation": 0,
  "opacity": 1,
  "fontColor": "#FFFFFF",
  "fontSize": 96,
  "textAlign": "center",
  "fontWeight": 700,
  "lineHeight": 1.4,
  "zIndex": 5,

  "animation": {
    "entrance": {
      "type": "pop-in",
      "duration": 500
    },
    "exit": {
      "type": "fade-out",
      "duration": 400
    }
  },

  "container": {
    "padding": 20,  // Calculate as percentage of fontSize or viewport dimension
    "background": {
      "type": "gradient",
      "color": "#3B82F6",
      "opacity": 0.9,
      "gradient": {
        "from": "#3B82F6",
        "to": "#8B5CF6",
        "direction": "to-right"
      }
    },
    "border": {
      "radius": 12,  // Calculate as percentage of viewport dimension
      "color": "#FFFFFF",
      "width": 2  // Calculate as percentage of viewport dimension
    },
    "backdropBlur": "md"
  }
}

平台分布

Claude Code

27.81%
按下载量换算36

Codex

21.29%
按下载量换算27

trae

17.33%
按下载量换算22

windsurf

11.76%
按下载量换算15

Antigravity

7.02%
按下载量换算9

github-copilot

2.88%
按下载量换算4

安全审计

Gen Agent Trust Hub

可疑

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。

继续浏览同类 Skills