{"id":"5158edef-5bf3-4d36-9883-7f578dac7509","slug":"gf691271-openclaw-link-capture","name":"link-capture","description":"链接知识捕获器。Frank（或任何agent）对话中出现URL时自动触发： 抓取全文→摘要→打标签→去重检查→存入Nowledge Mem知识图谱。 支持Twitter/X、网页文章、YouTube（摘要）。 附带选题卡功能：关键词→搜知识库+社媒热度+历史去重→输出完整选题卡。 激活词：任何URL（自动）；「选题 [关键词]」「选题卡 [关键词]」（手动）","canonicalUrl":"https://www.xpersona.co/skill/gf691271-openclaw-link-capture","sourceUrl":"https://github.com/gf691271/openclaw-link-capture","homepage":null,"source":"GITHUB_OPENCLEW","vendor":{"slug":"gf691271","label":"Gf691271","url":"https://github.com/gf691271/openclaw-link-capture"},"protocols":["OPENCLEW"],"capabilities":[],"trustScore":null,"trustConfidence":"unknown","artifactCount":0,"benchmarkCount":0,"lastRelease":null,"freshnessAt":"2026-03-01T06:06:00.473Z","freshnessLabel":"Mar 1, 2026","securityReviewed":true,"openapiReady":false,"stats":[{"label":"Trust score","value":"Unknown"},{"label":"Compatibility","value":"OpenClaw"},{"label":"Freshness","value":"Mar 1, 2026"},{"label":"Vendor","value":"Gf691271"},{"label":"Artifacts","value":"0"},{"label":"Benchmarks","value":"0"},{"label":"Last release","value":"Unpublished"}],"factsPreview":[{"factKey":"docs_crawl","category":"integration","label":"Crawlable docs","value":"6 indexed pages on the official domain","href":"https://github.com/login?return_to=https%3A%2F%2Fgithub.com%2Fopenclaw%2Fskills%2Ftree%2Fmain%2Fskills%2Fasleep123%2Fcaldav-calendar","sourceUrl":"https://github.com/login?return_to=https%3A%2F%2Fgithub.com%2Fopenclaw%2Fskills%2Ftree%2Fmain%2Fskills%2Fasleep123%2Fcaldav-calendar","sourceType":"search_document","confidence":"medium","observedAt":"2026-04-15T05:03:46.393Z","isPublic":true},{"factKey":"vendor","category":"vendor","label":"Vendor","value":"Gf691271","href":"https://github.com/gf691271/openclaw-link-capture","sourceUrl":"https://github.com/gf691271/openclaw-link-capture","sourceType":"profile","confidence":"medium","observedAt":"2026-03-01T06:06:01.859Z","isPublic":true},{"factKey":"protocols","category":"compatibility","label":"Protocol compatibility","value":"OpenClaw","href":"https://www.xpersona.co/api/v1/agents/gf691271-openclaw-link-capture/contract","sourceUrl":"https://www.xpersona.co/api/v1/agents/gf691271-openclaw-link-capture/contract","sourceType":"contract","confidence":"medium","observedAt":"2026-03-01T06:06:01.859Z","isPublic":true},{"factKey":"handshake_status","category":"security","label":"Handshake status","value":"UNKNOWN","href":"https://www.xpersona.co/api/v1/agents/gf691271-openclaw-link-capture/trust","sourceUrl":"https://www.xpersona.co/api/v1/agents/gf691271-openclaw-link-capture/trust","sourceType":"trust","confidence":"medium","observedAt":null,"isPublic":true}],"highlights":["Trust evidence available"],"agentCard":{"name":"link-capture","description":"链接知识捕获器。Frank（或任何agent）对话中出现URL时自动触发： 抓取全文→摘要→打标签→去重检查→存入Nowledge Mem知识图谱。 支持Twitter/X、网页文章、YouTube（摘要）。 附带选题卡功能：关键词→搜知识库+社媒热度+历史去重→输出完整选题卡。 激活词：任何URL（自动）；「选题 [关键词]」「选题卡 [关键词]」（手动）","source":"GITHUB_OPENCLEW","sourceId":"github:1166869565","repository":"https://github.com/gf691271/openclaw-link-capture","documentation":"https://www.xpersona.co/skill/gf691271-openclaw-link-capture/agent/gf691271-openclaw-link-capture","protocols":["OPENCLEW"],"languages":["typescript"],"install":{"command":"git clone https://github.com/gf691271/openclaw-link-capture.git","ecosystem":"git"},"examples":[{"kind":"example","language":"text","snippet":"https://x.com/...          → Twitter/X 推文\nhttps://twitter.com/...    → Twitter/X 推文  \nhttps://t.co/...           → Twitter短链（先解析再处理）\nhttps://youtube.com/...    → YouTube（抓字幕/描述，无字幕则存标题+描述）\nhttps://*.substack.com/... → Newsletter文章\nhttps://...（其他）        → 通用网页文章"},{"kind":"example","language":"text","snippet":"━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nSTEP A1 — 内容抓取\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nTwitter/X URL：\n  → 调用 x-tweet-fetcher: \n    python3 skills/x-tweet-fetcher/scripts/fetch_tweet.py --url \"<url>\" --pretty\n  → 提取：author, screen_name, text, likes, retweets, bookmarks, views, created_at\n\nYouTube URL：\n  → 调用 summarize skill（若已安装）\n  → 或 web_fetch 抓取页面获取标题+描述+自动字幕\n  → 无字幕则存：标题+频道+描述+URL\n\n其他网页：\n  → 调用 web_fetch(url, extractMode=\"markdown\")\n  → 截取前5000字（避免token浪费）\n  → 提取：标题、作者/来源、发布日期、核心内容\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nSTEP A2 — 去重检查\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n□ 用标题+核心句子做 memory_search（maxResults=3）\n□ 检查返回结果的相似度分数：\n  - score > 0.85：高度重复 → 告知已有，跳过存储，显示已有记录链接\n  - score 0.60-0.85：部分重复 → 告知相似内容，询问是否仍要存入（默认：存，但标注关联）\n  - score < 0.60：新内容 → 直接存入\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nSTEP A3 — 自动标签生成\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n根据内容自动推断标签（最多5个）：\n\n主题标签（选1-2）：\n  ai-agents | twitter | openclaw | real-estate | school-district\n  robotics | content-strategy | engineering | immigration | career\n  north-shore-crossing | zealty | dsr | arc | etl | family\n\n来源标签（选1）：\n  source-twitter | source-youtube | source-web | source-newsletter\n\n信号类型标签（选1）：\n  signal-tool | signal-insight | signal-news | signal-method | signal-data\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nSTEP A4 — 存入知识图谱\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n调用 nowledge_mem_save()：\n\ntitle: \"[来源] 标题（≤60字）\"\n  格式：\n    Twitter: \"[@screen_name] 推文核心观点（≤40字）\"\n    文章: \"文章标题 — 来源名\"\n    YouTube: \"[YouTube] 视频标题 — 频道名\"\n\ntext: 结构化摘要，格式：\n  来源：[URL / @作者 / 发布日期]\n  核心观点：[2-3句话，提炼最有价值的信息]\n  关键数据/金句：[数字、可引用的句子]\n  与Frank相关：[为什么这对Frank有用，1句话]\n  [如有代码/工具：工具名+用途]\n\nunit_type: 按内容选择\n  fact（工具/数据/事实）| learning（方法论/洞察）\n  event（新闻/发布）| preference（Frank明确表态喜欢的）\n\nlabels: [上一步生成的标签]\nimportance: \n  0.8-1.0：Frank主动分享+高互动（likes>1000 or bookmarks>500）\n  0.5-0.7：一般参考资料\n  0.3-0.4：背景信息\n\nevent_start: 内容发布日期（不是存储日期）\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nSTEP A5 — 回复确认（简洁）\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n存储成功后回复（1-3行，"}]}}