菜单
Collect top-N YouTube hot videos for any domain from the last 24 hours with Patchright, then output title, intro snippet, metrics, and link. Use when asked for a ranked trend digest in a specific field such as AI, finance, robotics, gaming, cybersecurity, or any other topic.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
基于 SOC 职业分类
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/yang0/skill_youtube_topN_subtitle --skill youtube-domain-topn-videos命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
| name | youtube-domain-topn-videos |
| description | Collect top-N YouTube hot videos for any domain from the last 24 hours with Patchright, then output title, intro snippet, metrics, and link. Use when asked for a ranked trend digest in a specific field such as AI, finance, robotics, gaming, cybersecurity, or any other topic. |
Use Patchright to crawl YouTube search results filtered to "Today", generate multi-query candidates from a chosen domain, rank by a recency-adjusted hot score, and export top-N items.
pip install patchright
python -m patchright install chromium
Default query templates are in references/query_templates.txt.
{domain} placeholder in each line.{domain} tutorial--queries-file to pass explicit queries.python scripts/fetch_topn_domain_videos.py --domain "robotics" --top 10 --hours 24 --cookies-file H:/cookies/youtube.txt
Useful options:
--domain <text>: Required domain/topic, for example cybersecurity or 跨境电商.--queries-file <path>: Use explicit query lines directly.--query-templates-file <path>: Use {domain} templates.--cookies-file <path>: Load YouTube cookies (Netscape or JSON).--max-results-per-query <n>: Increase candidate pool before ranking.--gl <country> and --hl <lang>: Control YouTube geo/language.--headed: Run non-headless for debugging.--output-json <path> and --output-md <path>: Change output files.Default output layout:
outputs/YYYY/MM/DD/YYYYMMDD_top{N}_{domain}_{hours}h.json and .mdEach ranked record includes:
title: Video title.intro: Description snippet from search card.data: Channel, views, publish text, age-hours estimate, duration, hot score, domain, matched queries.link: Canonical YouTube watch link.--scroll-rounds and --max-results-per-query.--gl explicitly (for example US, JP, TW).scripts/fetch_topn_domain_videos.py.