← Back to Skills Marketplace
xiatian5

跨平台内容自动发布

by xiatian5 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
105
Downloads
1
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install cross-platform-auto-poster
Description
跨平台内容自动化发布工作流,支持AI原创内容生成、视频制作、自动发布到小红书、抖音等平台,支持国内视频搬运去重后发布到TikTok/YouTube。
README (SKILL.md)

跨平台内容自动发布 📤🎬

全自动化跨平台内容发布工作流,支持AI原创内容生产 + 自动发布,也支持国内视频搬运去重后发布到海外平台。一键完成从选题到发布全流程,帮你轻松做内容变现。

触发词

当用户说这些话时,调用这个技能:

  • "跨平台自动发布"
  • "自动发布小红书"
  • "视频搬运自动化"
  • "TikTok自动发布"
  • "YouTube自动上传"
  • "内容自动化变现"
  • "小红书自动发文"

功能特性

功能 说明
AI原创小红书图文 从选题→文案→封面→发布全自动
AI生成口播视频 文案转语音+配图生成视频
视频搬运去重 自动下载→去重处理→改文案→发布
多平台支持 小红书、抖音、TikTok、YouTube
账号隔离 支持多账号管理,保存登录状态
数据统计 自动记录互动数据,生成报表

支持的工作流

工作流1: AI原创小红书图文 (推荐新手快速起步)

1. 热点选题 → 抓取小红书热搜,推荐热门选题
2. 内容生成 → 生成符合小红书风格的文案
3. 封面制作 → AI生成封面图
4. 自动发布 → 浏览器自动登录,上传图文发布
5. 数据记录 → 记录链接和发布时间

工作流2: AI原创小红书口播视频

1. 选题文案 → 同图文工作流
2. 语音合成 → 文案转自然语音
3. 配图生成 → 根据文案生成配图
4. 视频合成 → 语音+配图+字幕 → 输出视频
5. 自动发布 → 上传视频发布

工作流3: 国内视频搬运到TikTok/YouTube

1. 视频采集 → 按关键词下载热门视频
2. AI去重 → 裁切+翻转+调色+换BGM
3. 翻译改写 → 中文标题文案翻译成英文
4. 自动上传 → 发布到TikTok/YouTube

依赖

  • xiaohongshu-content-automation - 小红书内容生成
  • playwright-browser-automation - 浏览器自动化发布
  • image_generate - AI生成封面图
  • remotion-video-toolkit - 视频合成剪辑
  • openai-whisper-api - 语音识别(字幕生成)
  • ffmpeg - 视频处理(需要安装)

配置说明

首次使用

  1. 安装依赖 npm install 安装 Node 依赖
  2. 安装 playwright 浏览器 npx playwright install chromium
  3. 安装 ffmpeg 并添加到 PATH
  4. 配置账号信息在 config/accounts.json

账号配置模板

{
  "xiaohongshu": [
    {
      "name": "账号1",
      "cookiePath": "./auth/xiaohongshu-account1.json"
    }
  ],
  "tiktok": [],
  "youtube": []
}

使用示例

生成并发布一篇小红书图文

const workflow = require('./src/workflows/xiaohongshu-text-image');

await workflow.run({
  topic: "新手健身5个误区",
  style: "干货",
  accountIndex: 0
});

批量搬运视频到TikTok

const搬运工作流 = require('./src/workflows/video-cross-post');

await 搬运工作流.run({
  sourceKeyword: "fitness tips",
  sourcePlatform: "douyin",
  targetPlatform: "tiktok",
  count: 5
});

最佳实践

  1. 垂直领域 - 专注一个领域比如健身、美妆、美食,更容易做起来
  2. 批量生产 - 一次生成10-20篇,定时分批发布
  3. 数据驱动 - 看哪个选题互动好,多做同类内容
  4. 账号安全 - 一个账号不要发太频繁,避免风控

更新日志

v1.0.0 (2026-03-30)

  • 初始版本
  • 完整的AI原创小红书图文工作流
  • 支持多账号管理
  • 预留了视频搬运和多平台接口

相关技能


如果你觉得这个技能有用,请给它点个星,谢谢!⭐

Usage Guidance
This skill automates browser logins and saves session cookies locally, and it includes code to spoof/evade automation detection (navigator.webdriver override, geolocation). Before installing: (1) Understand that auth files saved under auth/ (storageState cookies) are sensitive—store them securely and remove them if you stop using the skill. (2) The repo advertises full scraping, AI-generation and downloading workflows but many functions are placeholders; you will need to provide or trust external tools/skills (yt-dlp, xiaohongshu-content-automation, remotion, TTS) to get full behavior. (3) Automating cross-posting of others' content may violate platform rules or copyright—ensure you have rights to repost. (4) If you proceed, review the publishers' automation scripts and config paths, and run them in a controlled environment (separate accounts) until you confirm behavior. If you want a safer alternative, use tools that rely on official platform APIs and OAuth rather than browser cookie reuse and UI automation.
Capability Analysis
Type: OpenClaw Skill Name: cross-platform-auto-poster Version: 1.0.0 The skill bundle implements automated social media posting by capturing and storing sensitive browser session states (cookies and authentication JSON) in a local 'auth/' directory (seen in tiktok-publisher.js, youtube-publisher.js, and xiaohongshu-publisher.js). It also utilizes child_process.spawn to execute shell commands via ffmpeg for video processing in video-cross-post.js. While these functions are necessary for the stated purpose of content automation, the management of plaintext session tokens and the use of shell execution represent a significant attack surface that could be exploited via prompt injection to exfiltrate credentials or execute unauthorized commands.
Capability Assessment
Purpose & Capability
The README/SKILL.md claim a full end-to-end pipeline (AI content generation, video downloading, de-duplication, cross‑posting). However many core pieces in the shipped code are placeholders/stubs (e.g., generateContent returns empty content, downloadVideos returns [] in src/workflows/video-cross-post.js). The project also relies on external skills/peerDependencies (xiaohongshu-content-automation, playwright-browser-automation, remotion-video-toolkit) rather than implementing them. This mismatch between claimed functionality and delivered implementation is significant: the skill advertises capabilities it doesn't implement locally and expects other components to provide them.
Instruction Scope
Runtime instructions direct the agent/user to install Playwright and ffmpeg and to configure cookie files (config/accounts.json). The code reads/writes local cookie/storageState files under auth/ and writes logs to output/. Publishers use Playwright to navigate platform sites and save storageState (cookies). The publishers also add evasive automation changes (navigator.webdriver override in xiaohongshu-publisher.js) and set geolocation/permissions in tiktok-publisher.js — behavior intended to avoid detection and to present a US location. Those behaviors are coherent with cross-posting purpose but expand scope to include actions that can affect account fingerprinting and location spoofing. Also, SKILL.md instructs video downloading/scraping but the implementation for downloading is missing, so actual runtime behavior depends on additional tools/code not present here.
Install Mechanism
This is instruction-only with code files included; there is no install script that fetches remote archives. package.json has no runtime dependencies but lists peerDependencies. The SKILL.md asks users to run npm install and npx playwright install and to install ffmpeg manually — standard for this type of tool. No external arbitrary download URLs or opaque installers are present in the repo.
Credentials
The skill declares no required environment variables or credentials, which aligns with what is shipped. However it requires storing and reading browser storageState files (cookies) under auth/ and config/accounts.json; those files contain session tokens and should be treated as sensitive. The code will save authentication state (context.storageState) to user-specified cookiePath and read it back to automate logins. Requesting cookie files (not env secrets) is proportionate to automated publishing, but it creates a sensitive local credential surface that users must protect.
Persistence & Privilege
The skill does not request always:true and does not modify other skills or system-wide settings. It writes its own auth/ and output/ files (saving session state and logs), which is expected for this functionality. Autonomous invocation is allowed (default) but not combined with any exceptional privileges.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install cross-platform-auto-poster
  3. After installation, invoke the skill by name or use /cross-platform-auto-poster
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release: 支持AI原创小红书图文/视频、国内视频搬运到TikTok/YouTube
Metadata
Slug cross-platform-auto-poster
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is 跨平台内容自动发布?

跨平台内容自动化发布工作流,支持AI原创内容生成、视频制作、自动发布到小红书、抖音等平台,支持国内视频搬运去重后发布到TikTok/YouTube。 It is an AI Agent Skill for Claude Code / OpenClaw, with 105 downloads so far.

How do I install 跨平台内容自动发布?

Run "/install cross-platform-auto-poster" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is 跨平台内容自动发布 free?

Yes, 跨平台内容自动发布 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does 跨平台内容自动发布 support?

跨平台内容自动发布 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created 跨平台内容自动发布?

It is built and maintained by xiatian5 (@xiatian5); the current version is v1.0.0.

💬 Comments