专业的爆款短视频文案创作工具。通过对标抖音爆款视频,智能提取视频内容,深度拆解爆款因素,并结合用户需求创作出符合爆款规律的新文案。适用于短视频创作者、运营人员提升内容质量。
npx skills add https://github.com/anbeime/skill --skill viral-video-copywriting
yt-dlp>=2024.1.0
beautifulsoup4>=4.12.0
lxml>=4.9.0
requests>=2.31.0
两种输入方式供用户选择:
用户提供抖音视频链接后,调用脚本提取视频信息:
python3 /workspace/projects/viral-video-copywriting/scripts/extract_douyin_video.py <抖音视频URL>
脚本可提取的信息(如成功):
重要说明:
当自动提取失败或用户更方便时,直接提供以下信息:
智能体将基于这些信息进行后续分析。
智能体基于第一步获取的视频信息,站在专业短视频创作者和运营专家角度,进行全方位拆解分析。
关键目标:提取可学习的方法论,而不是总结内容本身
分析维度(参考 references/analysis-framework.md):
输出格式:
智能体将生成一份详细的《爆款方法论拆解报告》,重点提取可学习的技巧、框架、规律,而不是内容本身。
智能体引导用户描述自己的创作想法:
需要了解的信息:
智能体会主动提出澄清问题,确保创作方向准确。
智能体结合对标视频的爆款规律和用户的原创内容,创作新文案。
核心创作原则(最重要):
本 Skill 的核心价值是学习方法论,而不是复制内容,但文案长度必须严格对标。智能体必须:
创作流程:
输出格式:
智能体将提供2-3版不同风格的文案备选,每版包含:
完整文案内容:
[开头钩子:约X字,预计X秒]
[中间内容:约X字,预计X秒]
[结尾总结:约X字,预计X秒]
总计:约X字,预计X秒
对标参数对比:
| 参数 | 对标文案 | 新文案 | 偏差 | 要求 |
|------|---------|--------|------|------|
| 总字数 | X字 | X字 | X% | ±5%以内 |
| 开头字数 | X字 | X字 | X% | ±5%以内 |
| 中间字数 | X字 | X字 | X% | ±5%以内 |
| 结尾字数 | X字 | X字 | X% | ±5%以内 |
| 预计时长 | X秒 | X秒 | X% | ±5%以内 |
| 结构类型 | XXX | 学习对标 | 一致 | 借鉴框架 |
| 语言风格 | XXX | 学习对标 | 一致 | 借鉴风格 |
| 句式特征 | XXX | 学习对标 | 一致 | 借鉴技法 |
原创性声明:
设计思路说明:
亮点标注:
对标元素迁移:
建议的拍摄/表现方式:
如果用户对创作结果有反馈,智能体支持多轮迭代优化:
用户输入:
我创作一个知识科普类的短视频,这是对标视频的链接:
https://www.douyin.com/video/7300000000000000000
我的想法是讲"如何提高工作效率",目标受众是职场新人
第一步:获取视频信息
第二步:深度拆解分析
智能体生成分析报告:
第三步:了解用户需求
第四步:创作新文案
智能体提供3版文案备选(详细格式见上文)
使用 references/sample-video-data.md 中的示例数据,可以完整测试整个流程:
# 测试脚本提取功能
python3 /workspace/projects/viral-video-copywriting/scripts/extract_douyin_video.py "https://www.douyin.com/video/7300000000000000000"
Extract cognitive patterns and thinking fingerprints from any text. Use this skill when the user wants to analyze how someone thinks, understand cognitive style, profile writing or speech patterns, compare thinking styles between people, asks "what's my thinking style", "analyze how this person reasons", "cognitive profile", "thinking pattern", "DHDNA", "digital DNA", or wants to understand the mind behind any text. Also trigger when the user provides text and wants deeper insight into the author's reasoning patterns, decision-making style, or cognitive signature.
GSAP animation reference for HyperFrames. Covers gsap.to(), from(), fromTo(), easing, stagger, defaults, timelines (gsap.timeline(), position parameter, labels, nesting, playback), and performance (transforms, will-change, quickTo). Use when writing GSAP animations in HyperFrames compositions.
配图助手 - 把文章/模块内容转成统一风格、少字高可读的 16:9 信息图提示词;先定“需要几张图+每张讲什么”,再压缩文案与隐喻,最后输出可直接复制的生图提示词并迭代。
| YouTube clip generation and editing with automated workflows — pull source video, slice highlights, add captions, and export.
Best practices for writing Remotion animations that stay intuitive for agents and editable in Remotion Studio Visual Mode.
YouTube transcript extraction and content reformatting: given a YouTube video URL, opens the video's transcript panel, extracts all timestamped segments, and transforms the raw transcript into summaries, chapter outlines, Twitter/X threads, blog posts, or notable quotes. Use when the user shares a YouTube URL or video link, asks to summarize a video, get a transcript, extract content from a YouTube video, get YouTube captions, extract YouTube captions, download YouTube captions, transcribe YouTube video, YouTube video to text, make a thread from YouTube, YouTube to blog post, YouTube to article, pull transcript from YouTube, YouTube content extraction, convert YouTube to text, video to transcript. Also applies when user wants to reformat any YouTube video content into structured output (chapters, threads, blog articles, key quotes).
Create a professional, eye-catching blog post header image sized for web (1200×628) with optional title composition guidance.
Direct high-fidelity cinematic video with AI — translates creative intent into technical cinematographic directives for Veo3, Kling, and Luma video models via muapi.ai
Take anbeime/viral-video-copywriting from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.