TypingMindTypingMindThe best frontend for AI
Yyl Benchmark Breakdown logo

Yyl Benchmark Breakdown

Community
ttfake92-lab
yyl-benchmark-breakdown

拆解对标账号 / 竞品内容——丢一个链接进来,自动识别平台(抖音/小红书/B站/YouTube/公众号)和粒度(单条 or 整个账号),抓取内容后输出三件套:可复用爆款公式、画面+口播逐段拆解、人设与内容定位,并自动存档到 benchmarks/。短视频自动下载 + 转写口播 + 抽视觉帧(contact sheet 多模态读图),不只看 caption。当用户说「拆解这个账号/视频/笔记」「分析对标/竞品」「这条为什么火」「给链接拆内容」「teardown」「看看别人怎么做的」时使用。抓不到内容会引导粘贴文案/字幕,绝不直接报错。

Overview

Publisherttfake92-lab
Repositoryskills
Skill nameyyl-benchmark-breakdown
Stars
308
Forks
45
Bundled files
8
Links
Skill sourceRepository
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 8 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by ttfake92-lab on GitHub. Read the source before you install it.

Installation

Install the Yyl Benchmark Breakdown AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/ttfake92-lab/skills/tree/main/skills/yyl-benchmark-breakdown
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/ttfake92-lab/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/skills/yyl-benchmark-breakdown .claude/skills/yyl-benchmark-breakdown
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Yyl Benchmark Breakdown in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Yyl Benchmark Breakdown on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Yyl Benchmark Breakdown is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

YYL 对标账号拆解 Skill

做一件事:给一个链接,把对方的内容拆到能复用的程度。

产出固定三件套(见 references/output-template.md):

  1. 可复用爆款公式 —— 提炼成能直接套用的选题套路 + 结构模板 + 视觉公式。
  2. 画面 + 口播逐段拆解 —— 时间轴对齐,逐段标注画面/口播/钩子/节奏/转折/CTA + 两者协同方式。
  3. 人设与内容定位 —— 人设标签、视觉符号、内容矩阵、差异化打法。

这是分析型 skill,产出 markdown 报告,不产图。要做封面用 yyl-video-thumbnail,要做小红书图文用 yuyile-social-card-skill。

工作流(前置清单通过后自动开干)

  1. 前置清单(每次开工先跑,但不要反复打扰用户):

    • 跑 bash scripts/check-deps.sh。
    • 如果自检失败:不要开始拆解;先帮用户补环境。按自检输出说明缺什么、为什么需要、下一步命令是什么。
      • Docker 未安装/未启动 → 说明 Docker 是本地下载 API 的运行环境,引导用户安装并打开 Docker Desktop。
      • Douyin_TikTok_Download_API 或 XHS-Downloader API 不可达 → Docker 可用时运行 bash scripts/bootstrap-local-apis.sh,再重跑 bash scripts/check-deps.sh。
      • ffmpeg / whisper 缺失 → 按 INSTALL.md 给出安装命令,装好再重跑自检。
    • 如果必需依赖全部通过:不要问用户确认,直接进入第 1 步。
    • TikHub 是可选项:首次 setup 或目标平台需要 TikHub(YouTube/快手/海外平台/本地 API 失败)且没有 TIKHUB_API_KEY 时,给用户两个选择:「现在配置」或「先跳过」。跳过后继续处理本地 API 能覆盖的平台;不要因为 TikHub 缺失阻塞抖音/B站/TikTok/小红书单条。

    前置清单:

    • 必需:Docker、Douyin_TikTok_Download_API、XHS-Downloader API、ffmpeg、openai-whisper。
    • 可选:TikHub API key、Jina Reader key。

  2. 识别:从 URL 判断平台 + 粒度(单条/账号)。规则见 references/fetch-playbook.md。

    • 拿不准粒度就先按单条处理,抓到页面再修正。

  3. 准备证据包(优先跑自动流程):

    • 先跑 bash scripts/prepare-assets.sh "<url>"。
    • 它会自动创建工作目录,保存 metadata.json、下载视频/图片、转写 audio.txt、抽帧并生成 sheet_*.jpg / xhs_sheet_*.jpg。
    • 跑完先读工作目录里的 manifest.txt,回显一句「我抓到了什么」(平台/字段/缺什么),再进入拆解。

  4. 取数回退(自动流程失败时继续走,不要在第一级失败就停)。细节、端点、命令见 references/fetch-playbook.md。

    • ① 平台本地 API(主力):
      • 抖音 / TikTok / Bilibili → Douyin_TikTok_Download_API:curl http://localhost:80/api/hybrid/video_data?url=...
      • 小红书单条笔记 → XHS-Downloader:POST ${XHS_API_BASE:-http://127.0.0.1:5556}/xhs/detail。download:true 会下载图文/视频文件到 ${XHS_DOWNLOAD_DIR:-$HOME/.xhs-downloader/Volume/Download}。
    • ② TikHub(备用):YouTube、快手、海外平台、小红书本地 API 失败时用。需要 TIKHUB_API_KEY 环境变量。
    • ③ Jina Reader:公众号文章、普通网页、纯图文兜底,curl https://r.jina.ai/<url>。
    • ④ 引导粘贴:还拿不到 → 让用户发文案/字幕/截图,不报错。

  5. 短视频必做:下载 + 转写口播 + 抽视觉帧(三样都缺不可)

    • 抖音 / B站 / TikTok / YouTube 单条 → 用 download_addr 下视频 → ffmpeg 抽音轨 → whisper --model medium --language Chinese 转写 → ffmpeg 场景检测 + 节奏采样 + 拼 contact sheet。
    • 小红书视频笔记 → XHS-Downloader 先下载 mp4 → 复用同一套 ffmpeg + whisper + contact sheet 流程。
    • 抽完帧用 Read 工具把 /tmp/bm/sheet_*.jpg 喂给自己看——多模态读图是视觉拆解的核心证据。
    • 小红书图文笔记 → 用 XHS-Downloader 下载图片,按图片序号拼 contact sheet,把图序当作"视觉脚本"看;没有口播就明确标「无口播」,不硬编。
    • 公众号文章 → 跳过下载/转写,但正文和配图都要作为证据。
    • 命令模板见 references/fetch-playbook.md 第 3 节。

  6. 拆解:按通用四层 + 视觉七维 + 平台特化分析。

    • 脚本/结构/选题/转化 → references/breakdown-framework.md
    • 画面层(视觉钩子/景别/切点节奏/字幕/视觉符号/色调/场景调度)→ references/visual-framework.md
    • 单条 → 深拆这一条,画面与口播必须时间轴对齐(看协同/互补/错位/预告/强调)。
    • 账号 → 先抓 5–15 条样本,找重复出现的规律(选题、结构、视觉符号、发布、互动),再挑 1–2 条代表作深拆。

  7. 输出 + 存档:按 references/output-template.md 生成报告,同时存一份到 benchmarks/<平台>-<账号名>/<日期>-<标题slug>.md(目录不存在就建)。 存档是第二阶段「长期对标库」的地基,每次都存。

    • 短视频报告末尾必须附 ## 完整文字转写稿,逐字粘贴 audio.txt 的完整内容;不能只给摘要。
    • 图文/网页报告末尾必须附完整正文;如果确实无口播,明确写「无口播」。

  8. 收尾:交付前检查报告末尾是否已经附完整转写稿/正文,再给「这条/这个账号最值得我抄的 3 个点」+ 一句可执行的下一步建议。

铁律

  • 自检通过就自动开干:不要再问用户要不要开始;必须完成下载 → 转写 → 视觉识别 → 爆款拆解 → 存档。不能只停在证据包准备完成。
  • 永不空手而归:逐级回退走完;真抓不到就引导粘贴,绝不只甩一句「抓取失败」。
  • 短视频必转写 + 必读图:caption + 数据不够,口播 + 画面 contact sheet 两个证据都要,缺一拆不全。
  • 报告末尾必须附完整转写稿:短视频必须把 audio.txt 完整粘贴到 ## 完整文字转写稿;图文/网页附完整正文或标注「无口播」。这是交付硬门槛,不能只写摘要、不能漏。
  • 要可复用,不要复述:别复述对方讲了什么/拍了什么,要提炼为什么这么做、我怎么套用。
  • 基于证据:数据、原句来自实际抓取;抓不到的指标标「未获取」,不要编。
  • 第一性:每个结论回到「它解决了观众什么问题 / 戳了什么情绪」,不堆术语。
  • 始终存档到 benchmarks/,文件名含平台、账号、日期。

文件

  • INSTALL.md —— 依赖安装指南(给人看)+ 常见问题。
  • scripts/check-deps.sh —— 一键自检前置依赖,缺什么给装的命令。
  • scripts/bootstrap-local-apis.sh —— 一键拉起本地 Douyin / XHS 下载 API;Docker 缺失时给安装引导。
  • scripts/prepare-assets.sh —— 自动证据包流程:下载 → 转写 → 抽帧/拼图 → 生成 manifest。
  • references/fetch-playbook.md —— ★开工先读:平台/粒度识别表、四级取数回退、下载+转写+抽帧命令、粘贴引导话术。
  • references/breakdown-framework.md —— 通用四层拆解 + 各平台特化(脚本/结构层) + 爆款公式提炼法。
  • references/visual-framework.md —— ★短视频必读:画面七维拆解 + 视觉/口播时间轴对齐 + 视觉公式提炼法。
  • references/output-template.md —— 三件套交付模板(已含画面+口播对齐表) + 存档命名约定。

第二阶段(暂未实现,已留地基)

  • 长期对标账号库:基于 benchmarks/ 里沉淀的多份拆解,做跨账号/跨时间的规律汇总(谁的哪类选题稳定爆、行业共性、可追踪趋势)。
  • 长视频转写优化:>15 分钟视频自动切片 + 并行转写;或接商业 API(飞书妙记/通义听悟)。
  • 小红书账号级自动化:接 MediaCrawler 自托管,用于账号主页采样、评论与搜索;单条下载已经由 XHS-Downloader 覆盖。

等单链接拆解打磨稳了再做。

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Yyl Benchmark Breakdown AI skill do?

拆解对标账号 / 竞品内容——丢一个链接进来,自动识别平台(抖音/小红书/B站/YouTube/公众号)和粒度(单条 or 整个账号),抓取内容后输出三件套:可复用爆款公式、画面+口播逐段拆解、人设与内容定位,并自动存档到 benchmarks/。短视频自动下载 + 转写口播 + 抽视觉帧(contact sheet 多模态读图),不只看 caption。当用户说「拆解这个账号/视频/笔记」「分析对标/竞品」「这条为什么火」「给链接拆内容」「teardown」「看看别人怎么做的」时使用。抓不到内容会引导粘贴文案/字幕,绝不直接报错。

Why use Yyl Benchmark Breakdown on TypingMind?

Because you install it once and use it with any model. Yyl Benchmark Breakdown is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Yyl Benchmark Breakdown in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/ttfake92-lab/skills/tree/main/skills/yyl-benchmark-breakdown. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Yyl Benchmark Breakdown?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Yyl Benchmark Breakdown?

As many as you like. As long as a model supports skills, you can use Yyl Benchmark Breakdown with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Yyl Benchmark Breakdown AI skill free?

It is published on GitHub by ttfake92-lab. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

Related AI Skills

View all
openclaw logo

Proof Video

openclaw
OrganizationPopular

Add subtitles, captions, narration cues, or zoom to a proof video or PR recording using repo-local capture helpers and a system ffmpeg renderer.

82.2K
391K
Instructions
View
openclaw logo

Facetime

openclaw
OrganizationPopular

Inspect FaceTime internal stages, place an approved call to a configured owner handle, and request hangup through facetime_call.

82.2K
391K
1 files
View
openclaw logo

Camsnap

openclaw
OrganizationPopular

Capture frames or clips from RTSP/ONVIF cameras and local webcams, including USB pan/tilt/zoom control.

82.2K
391K
Instructions
View
affaan-m logo

Content Engine

affaan-m
CommunityPopular

Create platform-native content systems for X, LinkedIn, TikTok, YouTube, newsletters, and repurposed multi-platform campaigns. Use when the user wants social posts, threads, scripts, content…

40.4K
270.2K
1 files
View

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇