YouTube Video Summarizer (Xiaolongxia)

总结 YouTube 视频内容，自动获取视频信息、搜索相关报道、生成结构化详细总结。支持中英文输出。

This skill looks internally consistent with its purpose: it fetches YouTube pages and transcripts, searches for related news, and formats structured summaries. Things to consider before installing: (1) network access — the agent will make outbound requests (YouTube pages, article fetches, web search) so expect telemetry and external HTTP traffic; (2) local execution — node scripts included will run locally if invoked, so review them if you have strict execution policies; (3) optional persistence — the SKILL.md mentions saving summaries to memory/YYYY-MM-DD.md and an EvoMap publish helper is included (the provided publish scripts only print data and do not send it, but they could be modified), so confirm whether you want summaries persisted or shared; (4) private or copyrighted videos — private/deleted videos cannot be fetched and transcripts may be unavailable. If you have concerns, run the scripts in an isolated environment, review the code, and test with non-sensitive video URLs before enabling automatic behaviors.

Type: OpenClaw Skill Name: xiaolongxia-youtube-summarizer Version: 1.1.0 The skill bundle is designed for YouTube video summarization. The `SKILL.md` provides clear, task-oriented instructions for the AI agent, utilizing `web_fetch` and `web_search` for legitimate content gathering from YouTube and related news sources. The `scripts/youtube-info.js` helper script performs web scraping on YouTube, confining its network requests to YouTube domains for video metadata and transcripts. The `evomap-publish.js` and `publish-to-evomap.js` scripts generate JSON payloads containing metadata about the skill itself (including `env_fingerprint` of the Node.js environment), but they only print this JSON to standard output and do not initiate any network requests to exfiltrate or publish this data externally. There is no evidence of intentional harmful behavior such as credential theft, unauthorized data exfiltration, persistence mechanisms, or malicious prompt injection.