Token导航 LogoToken导航

Agent Skill · 效率

shellbot-video-generator(shellbot 视频生成器)

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。

shellbot-video-generator 是效率类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描有提醒项,按安装量在效率类的 21,003 个 Skill 里排第 6,642。

总安装 9,056周安装 389GitHub Stars 1下载 3,174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-0
版本
1.0.0

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:shellbot-video-generator(shellbot 视频生成器)
来源仓库:https://github.com/cohnen/shellbot-video-generator
安装命令:
openclaw skills install shellbot-video-generator
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。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install shellbot-video-generator

这个 Skill 是做什么的

效率类 Skill 把重复的整理、汇总、转换工作交给助手完成,减少手工复制粘贴。

Shellbot-video-generator 基于 Remotion 提供 AI 驱动的视频制作工作流程。适用于广告、短片和动态图形等内容创作。通过 clawhub 安装,安装命令为 openclaw skills install shellbot-video-generator。需指定视频规格如分辨率与帧率。

建议预留足够渲染时间避免超时中断。

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 OpenClaw;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 openclaw skills install shellbot-video-generator;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

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

shellbot-video-generator 怎么安装?

在终端执行 openclaw skills install shellbot-video-generator。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

shellbot-video-generator 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。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。

shellbot-video-generator 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

shellbot-video-generator 是免费的吗?

来源标注的许可证是 MIT-0。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。

shellbot-video-generator 用的人多吗?

公开数据是累计安装 9,056 次,近一周安装 389 次,下载 3,174 次,GitHub 1 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。

shellbot-video-generator 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

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 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 OpenClaw,想直接扩展它能力的人。
  • 需要处理效率类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

效率类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
github

GitHub 代码协作 · OpenClaw

7,632,918545需要联网MIT-0通过
nano-pdf

Nano PDF 文档 · OpenClaw

4,249,476231需要联网MIT-0通过
auto-updater

自动更新程序 · OpenClaw

2,964,063386操作浏览器MIT-0通过
video-frames

视频帧 · OpenClaw

2,621,304112需要联网MIT-0通过
gemini

Gemini 助手 · OpenClaw

2,547,39648需要联网MIT-0有提醒
shellbot-video-generator

shellbot 视频生成器 · 本页

9,0561接触敏感数据MIT-0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

name
shellbot-video-generator
description
AI video production workflow using Remotion. Use when creating videos, short films, commercials, or motion graphics. Triggers on requests to make promotional videos, product demos, social media videos, animated explainers, or any programmatic video content. Produces polished motion graphics, not slideshows. Requires FIRECRAWL_API_KEY for website scraping and brand asset extraction.
metadata
openclaw
primaryEnv
FIRECRAWL_API_KEY
requires
env
homepage
https://getshell.ai

Video Generator (Remotion)

Create professional motion graphics videos programmatically with React and Remotion.

Non-Negotiable Story Structure (AIDA)

Every video MUST follow the AIDA arc in order:

  1. Attention — Establish the problem being solved. High-impact hook in the first 3 seconds.
  2. Interest — Introduce the solution. Show what it is and why it matters.
  3. Desire — Show use cases, not feature lists. Demonstrate real scenarios the viewer relates to.
  4. Action — End with a CTA that includes an incentive (discount, free trial, limited offer).

Default Workflow (ALWAYS follow this)

  1. Scrape brand data (if featuring a product) using Firecrawl
  2. Create the project from the local template:
   cd output && bash ../scripts/remotion.sh init <project-name>
  1. Install dependencies with npm install
  2. Build all scenes with proper motion graphics
  3. Start Remotion Studio as a background process:
   cd output/<project-name> && npm run dev

Wait for "Server ready" on port 3000.

  1. Expose via Cloudflare tunnel so user can access it:
   bash skills/cloudflare-tunnel/scripts/tunnel.sh start 3000
  1. Send the user the public URL (e.g. https://xxx.trycloudflare.com)

The user will preview in their browser, request changes, and you edit the source files. Remotion hot-reloads automatically.

Rendering (only when user explicitly asks to export):

cd output/<project-name>
npx remotion render CompositionName out/video.mp4

Quick Start

# Scaffold project from local template (no network needed)
cd output && bash ../scripts/remotion.sh init my-video
cd my-video && npm install

# Add motion libraries
npm install lucide-react

# Start dev server
npm run dev

# Expose publicly
bash skills/cloudflare-tunnel/scripts/tunnel.sh start 3000

Collecting Visual Assets

MANDATORY: Every video needs visual assets — logos, screenshots, product images, brand colors. Collect them BEFORE designing scenes. There are three sources:

1. Firecrawl Scrape (for any product/company URL)

Scrapes a website and extracts brand data + downloads reusable images in one step:

# Scrape and auto-download assets to the project's public/images/brand/
bash scripts/firecrawl.sh "https://example.com" output/my-video/public/images/brand

This returns JSON with: brandName, tagline, headline, description, features, logoUrl, faviconUrl, primaryColors (hex codes), ctaText, socialLinks, imageUrls (hero images, product shots, illustrations).

It also auto-downloads to the output directory: screenshot.png, og-image.png, favicon.png, logo.png, and all extracted page images (image-1.png, image-2.png, ...).

API Key: Set FIRECRAWL_API_KEY in the environment or .env.

2. User-Provided Assets

The user may provide logos, images, or screenshots directly (as file paths, URLs, or paste). Always ask if they have specific assets they want included. Save them to public/images/:

mkdir -p public/images
curl -sL "https://user-provided-url.com/logo.png" -o public/images/logo.png

3. Manual Download from Scrape Data

If you need specific images spotted in the scrape markdown (e.g. plan cards, channel icons), download them directly:

curl -sL "https://example.com/path/to/image.png" -o public/images/image.png

Using Assets in Remotion

Reference downloaded assets with staticFile():

import { Img, staticFile } from "remotion";

<Img src={staticFile("images/brand/logo.png")} />
<Img src={staticFile("images/brand/screenshot.png")} />

Core Architecture

Scene Management

Use scene-based architecture with proper transitions:

const SCENE_DURATIONS: Record<string, number> = {
  intro: 3000,     // 3s hook
  problem: 4000,   // 4s dramatic
  solution: 3500,  // 3.5s reveal
  features: 5000,  // 5s showcase
  cta: 3000,       // 3s close
};

Video Structure Pattern

import {
  AbsoluteFill, Sequence, useCurrentFrame,
  useVideoConfig, interpolate, spring,
  Img, staticFile, Audio,
} from "remotion";

export const MyVideo = () => {
  const frame = useCurrentFrame();
  const { fps, durationInFrames } = useVideoConfig();

  return (
    <AbsoluteFill>
      {/* Background music */}
      <Audio src={staticFile("audio/bg-music.mp3")} volume={0.35} />

      {/* Persistent background layer - OUTSIDE sequences */}
      <AnimatedBackground frame={frame} />

      {/* Scene sequences */}
      <Sequence from={0} durationInFrames={90}>
        <IntroScene />
      </Sequence>
      <Sequence from={90} durationInFrames={120}>
        <FeatureScene />
      </Sequence>
    </AbsoluteFill>
  );
};

Motion Graphics Principles

AVOID (Slideshow patterns)

  • Fading to black between scenes
  • Centered text on solid backgrounds
  • Same transition for everything
  • Linear/robotic animations
  • Static screens
  • slideLeft, slideRight, crossDissolve, fadeBlur presets
  • Emoji icons — NEVER use emoji, always use Lucide React icons

PURSUE (Motion graphics)

  • Overlapping transitions (next starts BEFORE current ends)
  • Layered compositions (background/midground/foreground)
  • Spring physics for organic motion
  • Varied timing (2-5s scenes, mixed rhythms)
  • Continuous visual elements across scenes
  • Custom transitions with clipPath, 3D transforms, morphs
  • Lucide React for ALL icons (npm install lucide-react) — never emoji

Transition Techniques

  1. Morph/Scale - Element scales up to fill screen, becomes next scene's background
  2. Wipe - Colored shape sweeps across, revealing next scene
  3. Zoom-through - Camera pushes into element, emerges into new scene
  4. Clip-path reveal - Circle/polygon grows from point to reveal
  5. Persistent anchor - One element stays while surroundings change
  6. Directional flow - Scene 1 exits right, Scene 2 enters from right
  7. Split/unfold - Screen divides, panels slide apart
  8. Perspective flip - Scene rotates on Y-axis in 3D

Animation Timing Reference

// Timing values (in seconds)
const timing = {
  micro: 0.1-0.2,      // Small shifts, subtle feedback
  snappy: 0.2-0.4,     // Element entrances, position changes
  standard: 0.5-0.8,   // Scene transitions, major reveals
  dramatic: 1.0-1.5,   // Hero moments, cinematic reveals
};

// Spring configs
const springs = {
  snappy: { stiffness: 400, damping: 30 },
  bouncy: { stiffness: 300, damping: 15 },
  smooth: { stiffness: 120, damping: 25 },
};

Visual Style Guidelines

Typography

  • One display font + one body font max
  • Massive headlines, tight tracking
  • Mix weights for hierarchy
  • Keep text SHORT (viewers can't pause)

Colors

  • Use brand colors from Firecrawl scrape as the primary palette — match the product's actual look
  • Avoid purple/indigo gradients unless the brand uses them or the user explicitly requests them
  • Simple, clean backgrounds are generally best — a single dark tone or subtle gradient beats layered textures
  • Intentional accent colors pulled from the brand

Layout

  • Use asymmetric layouts, off-center type
  • Edge-aligned elements create visual tension
  • Generous whitespace as design element
  • Use depth sparingly — a subtle backdrop blur or single gradient, not stacked textures

Remotion Essentials

Interpolation

const opacity = interpolate(frame, [0, 30], [0, 1], {
  extrapolateLeft: "clamp",
  extrapolateRight: "clamp"
});

const scale = spring({
  frame, fps,
  from: 0.8, to: 1,
  durationInFrames: 30,
  config: { damping: 12 }
});

Sequences with Overlap

<Sequence from={0} durationInFrames={100}>
  <Scene1 />
</Sequence>
<Sequence from={80} durationInFrames={100}>
  <Scene2 />
</Sequence>

Cross-Scene Continuity

Place persistent elements OUTSIDE Sequence blocks:

const PersistentShape = ({ currentScene }: { currentScene: number }) => {
  const positions = {
    0: { x: 100, y: 100, scale: 1, opacity: 0.3 },
    1: { x: 800, y: 200, scale: 2, opacity: 0.5 },
    2: { x: 400, y: 600, scale: 0.5, opacity: 1 },
  };

  return (
    <motion.div
      animate={positions[currentScene]}
      transition={{ duration: 0.8, ease: "easeInOut" }}
      className="absolute w-32 h-32 rounded-full bg-gradient-to-r from-coral to-orange"
    />
  );
};

Quality Tests

Before delivering, verify:

  • Mute test: Story follows visually without sound?
  • Squint test: Hierarchy visible when squinting?
  • Timing test: Motion feels natural, not robotic?
  • Consistency test: Similar elements behave similarly?
  • Slideshow test: Does NOT look like PowerPoint?
  • Loop test: Video loops smoothly back to start?

Implementation Steps

  1. Firecrawl brand scrape — If featuring a product, scrape its site first
  2. Director's treatment — Write vibe, camera style, emotional arc
  3. Visual direction — Colors, fonts, brand feel, animation style
  4. Scene breakdown — List every scene with description, duration, text, transitions
  5. Plan assets — User assets + generated images/videos + brand scrape assets
  6. Define durations — Vary pacing (2-3s punchy, 4-5s dramatic)
  7. Build persistent layer — Animated background outside scenes
  8. Build scenes — Each with enter/exit animations, 3-5 timed moments
  9. Open with hook — High-impact first scene
  10. Develop narrative — Content-driven middle scenes
  11. Strong ending — Intentional, resolved close
  12. Start Remotion Studio — npm run dev on port 3000
  13. Expose via tunnel — bash skills/cloudflare-tunnel/scripts/tunnel.sh start 3000
  14. Send user the public URL — They preview and request changes live
  15. Iterate — Edit source, hot-reload, repeat
  16. Render — Only when user says to export final video

File Structure

my-video/
├── src/
│   ├── Root.tsx              # Composition definitions
│   ├── index.ts              # Entry point
│   ├── index.css             # Global styles
│   ├── MyVideo.tsx           # Main video component
│   └── scenes/               # Scene components (optional)
├── public/
│   ├── images/
│   │   └── brand/            # Firecrawl-scraped assets
│   └── audio/                # Background music
├── remotion.config.ts
└── package.json

Common Components

See references/components.md for reusable:

  • Animated backgrounds
  • Terminal windows
  • Feature cards
  • Stats displays
  • CTA buttons
  • Text reveal animations

Tunnel Management

# Start tunnel (exposes port 3000 publicly)
bash skills/cloudflare-tunnel/scripts/tunnel.sh start 3000

# Check status
bash skills/cloudflare-tunnel/scripts/tunnel.sh status 3000

# List all tunnels
bash skills/cloudflare-tunnel/scripts/tunnel.sh list

# Stop tunnel
bash skills/cloudflare-tunnel/scripts/tunnel.sh stop 3000

平台分布

OpenClaw

92.34%
按下载量换算2,931

安全审计

VirusTotal

通过

ClawScan

可疑

Static analysis

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills