← Back to Skills Marketplace
keson1521

一键收藏抖音B站视频到你的飞书多维表格

by 黎可升 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
45
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install video-collector
Description
抖音/B站视频收藏入库——接收视频链接,自动抓取信息、生成总结、分类、入库飞书多维表格。触发词:视频入库 / 收藏视频 / 帮我整理这个视频。
README (SKILL.md)

视频收藏入库

接收抖音/B站视频链接 → 自动抓取信息 → 生成总结+分类 → 写入飞书多维表格。


配置说明

使用前需要在 skill 文件顶部填写你的飞书多维表格配置:

Base Token:  你的Base Token(格式:app_xxxxxxxxxxxx)
Table ID:    你的Table ID(格式:tblxxxxxxxxxxxx)
Base URL:    https://my.feishu.cn/base/你的BaseToken

如何获取:在飞书多维表格页面 URL 中提取,例如 https://my.feishu.cn/base/U7DRbtjO6aMlJcsYBGUcvah2nsf,则 Base Token 为 U7DRbtjO6aMlJcsYBGUcvah2nsf


字段映射(飞书多维表格)

建议按以下结构创建多维表格字段:

字段名 类型 说明
标题 文本 视频标题(主字段)
平台 单选 抖音 / B站
链接 文本 原始视频链接
UP主/博主 文本 UP主/博主名称
时长 文本 如 "12:34"
类型 单选 知识/娱乐/资讯/教程/其他
主题 多选 AI/产品/编程/设计/投资/生活/职场/其他
内容总结 文本 2-4句内容概括
添加日期 日期 格式 yyyy/MM/dd
标签 多选 已看/想看/反复看/待整理

流程

1. 判断平台

  • 抖音:douyin.com / v.douyin.com
  • B站:bilibili.com / b23.tv

2. 抓取视频信息

统一方案:dokobot(支持抖音、B站及所有 JS 渲染页面)

# 用 dokobot 读取页面,自动 JS 渲染,所有平台统一处理
dokobot read "\x3Curl>" --local --timeout 5000

从返回内容中提取:

  • 标题:取第一行标题(去除链接和噪音)
  • UP主/博主:找 "UP主名" 或 "作者" 相关行
  • 时长:找 "mm:ss" 或 "时长" 相关行
  • 描述/简介:找正文描述段落
  • 发布时间:找 "发布于" 或日期格式内容
  • 点赞/播放:找数字+单位(万/千)

抖音短链补全:如果是 v.douyin.com/xxx 短链,先用 curl 获取完整 URL:

# 抖音短链 → 完整 URL
FULL_URL=$(curl -sL "https://v.douyin.com/xxx" -w "%{url_effective}" -o /dev/null)
# 然后送入 dokobot
dokobot read "$FULL_URL" --local --timeout 5000

3. 生成字段

类型判断(根据内容推断):

  • 教程/操作演示 → "教程"
  • 知识科普/深度解读 → "知识"
  • 资讯/热点/新闻 → "资讯"
  • 搞笑/娱乐/休闲 → "娱乐"
  • 无法判断 → "其他"

主题判断(多选,最多选3个):

  • 涉及 AI/大模型/ChatGPT/编程/模型训练 → ["AI", "编程"]
  • 涉及 产品经理/需求/Aha Moment/用户增长 → ["产品"]
  • 涉及 UI/UX/设计/视觉/交互 → ["设计"]
  • 涉及 投资/理财/股票/黄金/赚钱 → ["投资"]
  • 涉及 职场/求职/面试/职场人际 → ["职场"]
  • 涉及 生活日常/美食/情感/家居 → ["生活"]
  • 无法判断 → ["其他"]

标签:默认填 "想看",如果用户说"已看完"或类似表述则填 "已看"

内容总结:根据视频标题+描述,用 2-4 句话概括核心内容。

4. 入库

# 添加日期用当天,格式 "YYYY-MM-DD HH:mm:ss"
ADD_DATE=$(date +"%Y-%m-%d %H:%M:%S")

# 写入飞书(替换 \x3CYOUR_BASE_TOKEN> 和 \x3CYOUR_TABLE_ID> 为实际值)
lark-cli base +record-upsert \
  --base-token \x3CYOUR_BASE_TOKEN> \
  --table-id \x3CYOUR_TABLE_ID> \
  --json @./_record.json

其中 _record.json 内容格式:

{
  "标题": "视频标题",
  "平台": "抖音",
  "链接": "https://...",
  "UP主/博主": "UP主名",
  "时长": "5:31",
  "类型": "教程",
  "主题": ["AI", "设计"],
  "内容总结": "内容概括...",
  "添加日期": "2026-05-04 10:30:00",
  "标签": ["想看"]
}

5. 输出确认

✅ 已入库视频收藏库
🔗 https://my.feishu.cn/base/\x3CYOUR_BASE_TOKEN>

📋 {标题}
🎯 {平台} · {类型} · {主题}
💬 {内容总结}

错误处理

  • 抓取失败 → 检查 URL 是否有效,确认 dokobot 浏览器扩展已连接
  • 入库失败 → 检查 Base Token 和 Table ID 是否正确配置
  • 字段类型不匹配 → 确认 JSON 中字段值类型与飞书字段类型一致
Usage Guidance
Review this skill before installing. The Feishu video-collection workflow is legitimate, but the declared tool permissions should be tightened, especially removing unused rm and node access. Also confirm that dokobot and lark-cli are trusted local tools and that the Feishu credentials or table permissions only allow writing to the intended collection table.
Capability Analysis
Type: OpenClaw Skill Name: video-collector Version: 1.0.0 The skill is a legitimate utility for collecting video metadata from Douyin and Bilibili and saving it to a Feishu (Lark) table. It uses standard tools like `curl` for URL resolution, `dokobot` for web scraping, and `lark-cli` for data insertion, all of which are consistent with its stated purpose in SKILL.md. While it requires sensitive Feishu API tokens, these are handled locally by the agent to interact with the user's own database, and there is no evidence of data exfiltration or malicious intent.
Capability Assessment
Purpose & Capability
The stated purpose—fetching Douyin/Bilibili video information, summarizing it, and adding it to a Feishu table—is coherent and disclosed.
Instruction Scope
The allowed tool scope includes arbitrary rm and node commands that are not used or justified by the documented workflow.
Install Mechanism
The skill has no install spec or declared binaries, but the instructions depend on dokobot and lark-cli being available.
Credentials
Network fetching and Feishu writes fit the purpose, but broad local shell permissions such as Bash(rm *) and Bash(node *) are disproportionate.
Persistence & Privilege
The skill persistently writes records to a configured Feishu multidimensional table, which is expected for the purpose but should be reviewed for workspace permissions.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install video-collector
  3. After installation, invoke the skill by name or use /video-collector
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release: collect and organize Douyin/Bilibili video links, auto-fetch info, summarize, categorize, and add to Feishu Bitable. - Supports video info extraction via dokobot for both platforms, handling short links and JS-rendered pages. - Includes configurable field mapping and recommended schema for Feishu Bitable. - Automated content type and topic detection, with default tags and summary generation. - Clear instructions for configuration, flow steps, and error handling.
Metadata
Slug video-collector
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is 一键收藏抖音B站视频到你的飞书多维表格?

抖音/B站视频收藏入库——接收视频链接,自动抓取信息、生成总结、分类、入库飞书多维表格。触发词:视频入库 / 收藏视频 / 帮我整理这个视频。 It is an AI Agent Skill for Claude Code / OpenClaw, with 45 downloads so far.

How do I install 一键收藏抖音B站视频到你的飞书多维表格?

Run "/install video-collector" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is 一键收藏抖音B站视频到你的飞书多维表格 free?

Yes, 一键收藏抖音B站视频到你的飞书多维表格 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does 一键收藏抖音B站视频到你的飞书多维表格 support?

一键收藏抖音B站视频到你的飞书多维表格 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created 一键收藏抖音B站视频到你的飞书多维表格?

It is built and maintained by 黎可升 (@keson1521); the current version is v1.0.0.

💬 Comments