Token导航 LogoToken导航TokenDH.com
研究检索external-servicegithub未标认证来源可访问许可证需确认审计提醒

yt-packagingyt 包装

Agent Skill

yt-packaging 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

222

周安装

9

GitHub Stars

公开资料未说明

下载量

70
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:yt-packaging(yt 包装)
来源仓库:https://github.com/naveedharri/benai-skills-private
仓库路径:skills/yt-packaging
安装命令:
npx skills add https://github.com/naveedharri/benai-skills-private --skill yt-packaging
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 npx skills 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。

skills.shnpx skills
npx skills add https://github.com/naveedharri/benai-skills-private --skill yt-packaging

简介

用于优化 YouTube 视频的分发与打包流程。

  • 支持转码设置、缩略图生成与元数据标准化。
  • 集成于自动化发布流水线中提升效率。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。
  • 需匹配目标平台编码规范与文件大小限制。
  • yt-packaging 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

YouTube Packaging

You are creating the title and thumbnail concept for a Ben AI YouTube video. Packaging is where CTR is determined — the title and thumbnail are the single biggest factor in whether a video succeeds.

Read references/youtube-strategy.md sections 2.1 (The One-Line Position), 2.3 (How We're Different), and 4.3 (Trending vs Evergreen) for positioning context.

Before You Start

You need from the user:

  1. The approved brief — Either load video-brief-{slug}.md from the working directory or get the brief details from the user
  2. Timing context — Is this a trending/update video (speed matters) or evergreen (optimize for long-term search)?
  3. Competitive context — What titles are competitors using for similar content? (Use WebSearch if not already known)

The Packaging Process

Step 1: Competitive Title Analysis

Before writing titles, research what's already out there:

  • Search YouTube for the topic
  • Note the top 5-10 existing titles
  • Identify patterns: what words/structures appear in high-view videos?
  • Identify gaps: what angle is NO ONE using?

Read references/title-formulas.md for proven title patterns.

Step 2: Generate Title Options

Create 5-10 title options across different strategies:

Strategy A: Direct Benefit Focus on what the viewer will be able to DO.

  • "How to [Verb] [Tool] to [Outcome]"
  • "Build [Thing] with [Tool] in [Time]"

Strategy B: Curiosity Gap Create intrigue without being misleading.

  • "The [Tool] Feature Nobody Is Talking About"
  • "I Automated My Entire [Process] — Here's How"

Strategy C: Authority/Definitive Position as THE resource.

  • "The Complete [Tool] Guide for [Year]"
  • "Everything You Need to Know About [Feature]"

Strategy D: Trending/Urgency For update videos and time-sensitive content.

  • "[Tool] Just Changed Everything — Here's What You Need to Know"
  • "New [Feature] Walkthrough: [Specific Capability]"

For each title:

  • Note which strategy it uses
  • Estimate search-friendliness (contains searchable keywords?)
  • Estimate curiosity factor (would you click?)
  • Note the character count (aim for under 60 characters for full display)

Step 3: Generate Thumbnail Concepts

Create 3-5 thumbnail concepts. Each concept should include:

Visual description:

  • Main visual element (screen-share preview, tool logo, face expression, graphic)
  • Text overlay (1-4 words MAX — the thumbnail is not the title)
  • Color scheme and contrast
  • Layout (rule of thirds, focal point)

Read references/thumbnail-principles.md for design guidelines.

For each concept:

  • How does it complement the title? (Title + thumbnail = one message, not two separate messages)
  • Does it stand out in a feed of similar videos?
  • Is it readable at mobile size (small thumbnail)?
  • Does it communicate the value proposition visually?

Step 4: A/B Test Variants

For the top 2-3 title + thumbnail combos, suggest A/B testing variants:

  • Same title, different thumbnail approach
  • Same thumbnail, different title strategy
  • Small tweaks (word swaps, number inclusion, emoji vs no emoji)

Step 5: Present to User

Present all options in a structured format:

## Title Options

| # | Title | Strategy | Keywords | Length |
|---|-------|----------|----------|--------|
| 1 | [title] | Direct Benefit | claude cowork, MCP | 48 chars |
| 2 | [title] | Curiosity Gap | claude, automation | 52 chars |
...

## Thumbnail Concepts

### Concept A: [Name]
- Visual: [description]
- Text overlay: "[text]"
- Complements titles: #1, #3

### Concept B: [Name]
- Visual: [description]
- Text overlay: "[text]"
- Complements titles: #2, #4

## Recommended Combo

Title #[X] + Thumbnail Concept [Y] because [reasoning]

"Pick your title and thumbnail direction."

  • Go with recommended combo
  • Mix and match (pick different title + thumbnail)
  • Request more options
  • Adjust the angle

This is a mandatory human checkpoint.

Step 6: Save

Save the approved packaging as packaging-{slug}.md.

Key Principles

  • Title and thumbnail are ONE system. They must work together. The thumbnail should NOT repeat the title — they should complement each other. Title says what, thumbnail shows the emotion/intrigue.
  • No clickbait without substance. Ben AI's brand is anti-hype. Titles should be compelling but honest. Every promise in the title must be delivered in the video.
  • Search + browse balance. For evergreen content, include searchable keywords. For trending content, prioritize curiosity and urgency.
  • Mobile-first thumbnails. Most YouTube browsing happens on mobile. Thumbnails must be readable at small sizes. Test by mentally shrinking the concept.
  • Under 60 characters. Titles get truncated on mobile. The most important words should be in the first 40 characters.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.09%
按下载量换算25

Claude

29.26%
按下载量换算20

Cursor

20.24%
按下载量换算14

Gemini CLI

9.59%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills