X Hot Topics Daily
每日追踪并推送 X(x.com)热点话题新闻简报。用于用户要求“每天定时看 X 热点”“按指定话题抓取热门帖”“用 browser 方式监控 X 趋势”时。默认覆盖 AI、LLM、社会热点(中国/新加坡/美国)五个主题。重点输出“最低成本 know-what 版”:2 条必知 + 3 条可忽略 + 10 分钟行动,帮助用户抗 FOMO,而不是只罗列标题。
MIT-0 · Free to use, modify, and redistribute. No attribution required.
⭐ 1 · 1.1k · 11 current installs · 11 all-time installs
by@hmzo
MIT-0
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The skill's stated purpose (daily X.com topic summaries) matches the instructions (open X, run topical searches, extract top posts). However the runtime explicitly asks to use 'browser profile=chrome', which implies access to a local Chrome profile or browser automation credentials; that access is not declared in the skill metadata (no required config paths or env vars). This is incongruent with the declared zero-credential footprint.
Instruction Scope
SKILL.md stays focused: open X, fetch top posts per topic, filter/ dedupe, and summarize. It does not instruct sending data to external endpoints beyond producing the summary. The main scope concern is the use of a browser snapshot (refs=aria) which can capture personalized content tied to a logged-in session if the Chrome profile is used.
Install Mechanism
Instruction-only skill with no install spec and no code files — lowest install risk. There is nothing being downloaded or written by the skill itself.
Credentials
The skill requests no environment variables or config paths, yet the instructions demand 'profile=chrome' browser access. That implicitly requires access to the user's browser environment or automation endpoint (cookies, session tokens, stored credentials). The lack of declared required config/permission is disproportionate and a privacy risk.
Persistence & Privilege
always:false and no persistent install behavior. The skill does not request system-wide configuration changes or permanent presence. Autonomous invocation is allowed (platform default) but not combined with other high privileges.
What to consider before installing
This skill appears to do what it says (scrape X and summarize), but it asks the agent to use a Chrome profile for browser automation without declaring that it will access browser data. That can expose your logged-in X account, cookies, and personalized feed. Before installing: 1) Confirm how your agent's 'browser' tool handles 'profile=chrome' — does it use your real profile or an isolated ephemeral profile? 2) If you don't want account-linked personalization or cookie exposure, require the skill to run with an unauthenticated/ephemeral profile or headless scraping. 3) Verify where snapshots and extracted data are stored or transmitted and the retention policy. 4) If you accept the privacy tradeoff, limit the agent's browser permissions or create a dedicated Chrome profile for this skill. If the publisher can provide explicit wording that no personal profiles or credentials are used (or supply a dedicated/profile path), that would reduce the concern.Like a lobster shell, security has layers — review code before you run it.
Current versionv1.0.0
Download ziplatest
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
SKILL.md
X Hot Topics Daily
核心目标
每天按固定话题抓取 X 热帖,产出“可快速决策”的中文简报:
- 只保留信息密度高、互动高、可行动的信息
- 去掉重复话题和低质量灌水
- 强制过滤常见营销诈骗与谣言型内容(送币抽奖、评论 done、无来源爆料等)
- 默认给出“最低成本 know-what”结论,减少 FOMO
固定工作流
1) 使用 browser 打开 X
优先使用 browser + profile=chrome:
browser action=open profile=chrome target=host targetUrl=https://x.com/explore/tabs/for_you
若无法连接浏览器控制服务,直接说明失败原因并提示重试,不改用无关工具链。
2) 按主题抓取(默认 5 个)
默认主题:AI、LLM、中国社会热点、新加坡社会热点、美国社会热点
建议查询页(Top):
- AI:
https://x.com/search?q=AI%20(lang%3Aen%20OR%20lang%3Azh)&src=typed_query&f=top - LLM:
https://x.com/search?q=LLM%20(lang%3Aen%20OR%20lang%3Azh)&src=typed_query&f=top - 中国社会热点:
https://x.com/search?q=(中国%20OR%20China)%20(社会%20OR%20热点%20OR%20新闻)%20(lang%3Azh%20OR%20lang%3Aen)&src=typed_query&f=top - 新加坡社会热点:
https://x.com/search?q=(Singapore%20OR%20新加坡)%20(news%20OR%20breaking%20OR%20热点)%20(lang%3Aen%20OR%20lang%3Azh)&src=typed_query&f=top - 美国社会热点:
https://x.com/search?q=(US%20OR%20United%20States%20OR%20美国)%20(news%20OR%20breaking%20OR%20热点)%20(lang%3Aen%20OR%20lang%3Azh)&src=typed_query&f=top
然后:
browser action=snapshot profile=chrome refs=aria- 从快照提取前排
article的:作者、发布时间、正文核心句、互动量(至少 likes/views) - 每个主题先抓 3-5 条候选,再筛到 1-2 条入选
3) 过滤与去重(强制)
先过滤再总结:
-
营销诈骗与低质过滤
- 丢弃包含以下模式的帖子:
giving away/random winner/like + repost + comment done- 夸张收益承诺、领奖导流、异常短链落地页
- 无可靠来源却宣称“突发/独家”的截图搬运
- 明显广告帖伪装成新闻或分析
- 丢弃包含以下模式的帖子:
-
去重
- 相同新闻事件跨多个主题重复出现时,只保留一次主条目
- 其余主题写“同一事件延伸讨论”一句话即可
-
新鲜度
- 优先过去 24 小时内容
- 若不足,再补充 48 小时内高价值帖并标注
-
区域平衡(社会热点专用)
- 中国/新加坡/美国三个区域尽量各保留 1 条有效信息
- 若某区域有效信息不足,明确写“该区域今天有效热点较少”
4) 内容提炼(强制,不可只列标题)
每条入选热点必须补齐以下信息:
- 发生了什么(1 句)
- 为什么值得现在关注(1 句,强调时效/影响)
- 谁最该关心(开发者/普通公众/投资者/出行者)
- 是否可执行(给“可做/可观望/可忽略”)
写法要求:
- 不复读标题,不堆链接
- 用“结论先行”写法,最多 3 句讲透
- 面向“最低成本 know-what”
5) 输出格式(默认:反 FOMO 速览)
默认按下列结构输出:
【X 热点追踪|最低成本 know-what】
日期:YYYY-MM-DD (Asia/Singapore)
主题:AI / LLM / 中国社会热点 / 新加坡社会热点 / 美国社会热点
A) 今天真正要知道的(2条必知)
- [事件1]:发生了什么;为什么现在重要;你该怎么理解
- [事件2]:发生了什么;为什么现在重要;你该怎么理解
B) 你可以直接忽略的(3条)
- [噪音类型/话题1]:忽略原因
- [噪音类型/话题2]:忽略原因
- [噪音类型/话题3]:忽略原因
C) 10分钟最低成本行动
- 1) 先看:...
- 2) 再看:...
- 3) 最后判断:...
D) 风险提醒
- 今日过滤掉的疑似诈骗/低质信息类型:...
兼容输出(用户明确要求时)
若用户明确说“按旧版日报格式”,再使用旧结构:
- 今日总览
- 分主题热点
- 风险提醒
- 给 hmzo 的一句建议
质量标准
- 全文中文,简洁但信息完整
- 每日总条目建议 5-10 条(最终只保留最有用信息)
- 不能输出“我无法保证准确”这类空话;要给明确结论
- 信息不足时明确写“今天该主题有效热点较少”
- 社会热点优先公共影响(政策、公共安全、交通民生、重大公共事件),避免娱乐八卦
- 重点是“可读、可判断、可行动”
- 若热点质量明显偏低,必须直说“今天噪音占比高,建议降低信息摄入”
快速触发语句
可用以下指令触发:
执行 x-hot-topics-daily,抓取今天 X 上 AI/LLM/中国/新加坡/美国社会热点并推送给我今天 X 五个主题的热点新闻速览(browser方式)给我反 FOMO 版本:2条必知+3条可忽略+10分钟行动
Files
1 totalSelect a file
Select a file to preview.
Comments
Loading comments…
