Token导航 LogoToken导航TokenDH.com
研究检索操作浏览器clawhub未标认证来源可访问clear审计提醒

ai-news-daily-new艾新闻日报新

Agent Skill

ai-news-daily-new 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

6,797

周安装

289

GitHub Stars

公开资料未说明

下载量

2,381
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:ai-news-daily-new(艾新闻日报新)
来源仓库:https://github.com/wangyieleven/ai-news-daily-new
安装命令:
openclaw skills install ai-news-daily-new
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install ai-news-daily-new

简介

从指定信源检索最近24小时AI新闻,完成去重/分类/摘要,生成结构化AI新闻日报。Invoke when user wants AI news digest or daily brief.

SKILL.md

name
ai-news-daily-brief
description
从指定信源检索最近24小时AI新闻,完成去重/分类/摘要,生成结构化AI新闻日报。Invoke when user wants AI news digest or daily brief.

AI News Daily Brief

Overview

This skill generates a structured daily AI news brief by collecting, deduplicating, classifying, and summarizing the most important AI news from the past 24 hours.

It is designed for users who want a fast, high-quality overview of recent AI developments without reading fragmented reports across many sources.

The skill focuses on a fixed set of trusted Chinese AI media sources, global technology media sources, and official first-party AI company sources. It produces a concise but information-dense daily brief containing 10–20 high-value news items, ranked and organized for easy reading.

This skill is especially useful for:

  • AI product managers
  • industry researchers
  • strategy teams
  • enterprise AI practitioners
  • users who want a daily AI news digest for tracking trends, competitors, and product updates

Output Language

Unless the user explicitly requests another language, all output must be written in Simplified Chinese.

Rules:

  1. Section headings, summaries, analysis, takeaways, and explanations must be written in Simplified Chinese.
  2. Original article titles may remain in their original language when necessary, but Chinese translation or Chinese summary should be provided.
  3. Do not output the final daily brief in English unless the user explicitly asks for English.
  4. Keep the writing concise, clear, and suitable for Chinese-speaking professional users.

Supported Sources

Chinese AI / tech media

  • AIBase(入口:https://news.aibase.com/zh/daily)
  • 36Kr AI
  • 机器之心
  • 量子位

Global media

  • Reuters
  • TechCrunch(新增)
  • The Information(新增)
  • The Verge
  • WIRED
  • VentureBeat

Official first-party sources

  • Google DeepMind Blog
  • OpenAI News
  • Anthropic Newsroom
  • Anthropic Engineering
  • NVIDIA Blog

Do not expand beyond these sources unless explicitly instructed by the user.


Search Language Rules

Use source-appropriate search language to improve recall and precision.

  1. For Chinese sources (AIBase, 36Kr AI, 机器之心, 量子位):

- Search primarily in Chinese. - Use Chinese topic phrases and Chinese summaries of events. - Preserve important English entity names when helpful, such as OpenAI, NVIDIA, DeepMind, Anthropic, Agent, RAG.

  1. For global and official sources (Reuters, The Verge, WIRED, VentureBeat, Google DeepMind Blog, OpenAI News, Anthropic Newsroom, Anthropic Engineering, NVIDIA Blog):

- Search primarily in English. - Use original English company names, product names, and technical terms. - When useful, map Chinese user intent into English search queries before searching.

  1. For major cross-language events:

- Try both Chinese and English query variants when necessary. - Merge duplicated results across languages into one unified story.

  1. Prefer bilingual keyword expansion for important entities and concepts when useful, for example:

- 英伟达 / NVIDIA - 谷歌DeepMind / Google DeepMind - 智能体 / Agent - 检索增强生成 / RAG - 推理模型 / reasoning model

  1. Final output language must still follow the output language rules:

- Default output is Simplified Chinese unless the user explicitly requests another language.


When to Use

Trigger this skill when the user wants:

  • an AI news daily brief
  • a daily digest of AI industry updates
  • a roundup of the latest AI news from the last 24 hours
  • a structured summary of important AI developments
  • a curated AI news report based on predefined trusted sources

Example user requests:

  • “帮我整理今天的AI新闻日报”
  • “汇总最近24小时AI圈的重要新闻”
  • “给我一份AI行业日报”
  • “基于固定信源整理今天的AI动态”
  • “Generate an AI news daily brief”

Core Task

Your job is to:

  1. Search the supported sources for AI-related news from the last 24 hours.
  2. Filter out low-value, weakly related, repetitive, or promotional content.
  3. Deduplicate overlapping reports about the same event.
  4. Prioritize official first-party announcements when available.
  5. Merge repeated coverage of the same event into a single unified news item.
  6. Rank the final news items by importance.
  7. Output a structured AI news daily brief. Standard target: 8–12 items (expand to 12–15 when news is dense; reduce to 5–8 when verified sources are limited).

Do not merely list headlines. Extract facts, summarize clearly, and explain why each item matters.


Time Scope

Default time window: the last 24 hours.

If an item is slightly older than 24 hours but remains highly relevant in the current cycle and is still a major topic across the supported sources, it may be included sparingly. If so, make that clear in the summary.


Hard Constraints

The skill must strictly follow these constraints:

  1. Only use the approved sources listed in references/sources.md.

- Do not use any source outside the approved whitelist unless the user explicitly requests broader coverage. - If a story cannot be traced back to an approved source, exclude it.

  1. Only include stories whose original publication time can be verified from the source page itself.

- The skill must open the source page and confirm the explicit original publication date/time. - Do not rely only on search snippets, search ranking, summaries, reposts, or secondary mentions.

  1. For the daily brief, only include stories published within the last 24 hours.

- If the original publication time is outside the last 24 hours, exclude the story. - If the publication time is ambiguous, missing, approximate, or cannot be confirmed, exclude the story.

  1. Never relax the time window, source whitelist, or factual verification rules just to increase the item count.

- If there are not enough verified high-value stories, return fewer items.

  1. Accuracy is more important than completeness.

- It is acceptable to output only 5–8 items if those are the only verified high-value stories in the time window.

  1. If a page is a landing page, archive page, category page, rolling update page, or release notes hub without a clearly verifiable article timestamp, do not use it as a final story source.

- Only use a specific article or post page with an explicit publication date/time. - If the article page cannot be opened by any extraction method (web_fetchbrowser.snapshot), exclude the story. Search snippets are not acceptable as sole evidence of publication time.

  1. Emergency exception — major breaking news: If a significant AI event is generating widespread coverage but ALL approved sources are unable to verify publication time (e.g., a major model launch breaking on a source not yet in the whitelist), you may include it as a "待核实" item with explicit note that time could not be verified. This exception requires: (a) the event is genuinely major and globally recognized; (b) at least one credible approved source links to or mentions the event; (c) the item is clearly labeled as "待核实". Do not use this exception to circumvent source quality rules for routine news.

Source Access — Tools & Fallback Chain

Different sources require different tools. Use the following recommended strategy:

Tools Comparison

ToolCoverageSpeedBest for
web_fetch9/14 sitesFastThe Verge, Wired, VentureBeat*, NVIDIA Blog, TechCrunch, The Information, Anthropic, OpenAI (article pages), AIBase, 36Kr
agent-browser2/14 sitesMedium机器之心, 量子位 (blocked by HTTP/Cloudflare)
tavily5/14 sitesFastReuters, DeepMind Blog, OpenAI (fallback), VentureBeat (primary to avoid 429), 机器之心 (fallback)

*VentureBeat triggers 429 rate limit via web_fetch; use tavily as primary

*OpenAI main page returns 403, but article-level URLs work via web_fetch

Recommended Extraction Order

  1. Try web_fetch first (fast, works for 9/14 sites: Western media + OpenAI article pages + AIBase + 36Kr)
  2. Try agent-browser for Chinese sites that block HTTP (机器之心, 量子位)
  3. Try tavily as primary for blocked domains (Reuters, DeepMind) and rate-limited sites (VentureBeat)
  4. Try article-level URLs — category/homepage blocked doesn't mean all article pages are

Site-by-Site Strategy

SourcePrimary ToolFallbackNotes
AIBaseweb_fetchagent-browserDirect fetch works
36Krweb_fetchagent-browserHomepage anti-bot; try section URLs
机器之心agent-browsertavilyHTTP blocked; browser required
量子位agent-browsertavilyCloudflare blocks; browser required
ReuterstavilyDomain blocked; tavily is the ONLY reliable path
TechCrunchweb_fetchtavilyDirect fetch works
The Informationweb_fetchtavilyDirect fetch works
The Vergeweb_fetchDirect fetch works
WIREDweb_fetchDirect fetch works
VentureBeattavilyagent-browserweb_fetch triggers 429 rate limit
DeepMind BlogtavilyRedirect loop; tavily is the ONLY reliable path
OpenAI Newsweb_fetch (article pages)tavilyHomepage returns 403; article URLs work
Anthropic Newsweb_fetchagent-browserDirect fetch works
NVIDIA Blogweb_fetchDirect fetch works

Strict Rule

If no method can access a source, the story must be excluded. Never pad with low-signal content just to hit a count target.


Inclusion Criteria

Include only news that is directly related to AI and has clear informational value.

Priority topics include:

  • model launches and major model updates
  • multimodal models, reasoning models, agent systems
  • AI products and feature launches
  • AI agents, workflow automation, AI office tools
  • enterprise AI, knowledge bases, RAG, copilots
  • chips, GPUs, inference infrastructure, data centers, compute
  • major partnerships, acquisitions, funding, or commercial milestones
  • AI policy, regulation, security, safety, copyright, governance
  • research breakthroughs with clear product or industry implications

Exclusion Criteria

Do not include:

  • weakly related general tech news
  • clickbait or headline-only articles
  • marketing copy or promotional articles
  • pure reposts with no additional information
  • speculative opinion pieces without factual basis
  • trivial updates with no clear industry significance
  • repeated reports of the same event as separate items

If a topic is not clearly AI-related or does not add meaningful signal, exclude it.


Source Priority Rules

When the same event appears across multiple sources, use the following priority order:

  1. Official first-party source
  2. Reuters
  3. TechCrunch / The Information / The Verge / WIRED / VentureBeat
  4. 机器之心 / 量子位 / 36Kr AI
  5. AIBase

If media interpretation differs from the official announcement, prefer the official source and note that other outlets framed it differently if needed.


Deduplication Rules

Before writing the brief, you must deduplicate and merge overlapping coverage.

Rules:

  • Treat the same event covered by multiple outlets as one story.
  • Keep the most authoritative and information-rich source as the primary source.
  • Record other relevant outlets as supplementary sources.
  • Write the news item around the event itself, not around any one outlet’s headline.
  • Do not repeat the same story in multiple categories.

Output Requirements

Prefer to keep 8–12 curated news items in the daily brief.

If there is an unusually dense and high-quality news cycle, the brief may expand to 12–15 items.

In exceptional cases with very high signal density, the brief may expand up to 20 items, but only when quality remains consistently high and all items pass verification.

If the number of genuinely high-value and fully verified stories is limited, the brief may be reduced to 5–8 items. Do not add filler, weakly related updates, approximate matches, or insufficiently verified stories just to reach a target count.

Always prioritize:

  • source whitelist compliance
  • publication time verification
  • factual accuracy
  • signal quality
  • readability

Output Format

AI新闻日报({date})

一、今日最重要的5条AI新闻

按重要性对 5 条新闻进行排序。

每条新闻包含:

  • 标题
  • 类别
  • 主来源
  • 发布时间
  • 事件摘要:100–150字
  • 关键看点:2–3条
  • 影响判断:50–100字,说明为什么值得关注
  • 原始链接
  • 补充来源(如有)

可用类别:

  • 模型与产品发布
  • Agent与应用落地
  • 企业与商业化动态
  • 基础设施 / 芯片 / 算力
  • 政策 / 监管 / AI安全
  • 研究进展与技术趋势

二、完整新闻清单(通常为12–15条,高价值不足时可降至8–10条)

按类别分组列出所有入选新闻。

每条包含:

  • 标题
  • 来源
  • 时间
  • 一句话摘要
  • 原始链接

类别顺序:

  1. 模型与产品发布
  2. Agent与应用落地
  3. 企业与商业化动态
  4. 基础设施 / 芯片 / 算力
  5. 政策 / 监管 / AI安全
  6. 研究进展与技术趋势

三、日报总结

最后给出:

  • 今日最值得关注的3个信号
  • 对AI产品经理最有价值的3条启示
  • 对企业AI / 知识库 / RAG / 办公场景最相关的机会点或风险点

Writing Style

  • Final output must be in Simplified Chinese by default.
  • Be concise, factual, and high-signal.
  • Do not copy headlines mechanically.
  • Compress information and explain it clearly.
  • Focus on facts first, then interpretation.
  • Avoid hype language and promotional phrasing.
  • Explicitly label uncertain claims as “待核实”.
  • Keep the brief readable in about 5 minutes.

Quality Checks

Before finalizing the output, verify:

  1. Duplicate stories have been merged.
  2. Weakly related or low-value items have been removed.
  3. Primary sources and supplementary sources are used appropriately, with approved official sources preferred when available.
  4. Every story includes a source and link.
  5. The top 5 are genuinely ranked by importance.
  6. The final brief contains only 10–20 meaningful items.
  7. The summaries explain not just what happened, but why it matters.
  8. The final output language is Simplified Chinese unless the user explicitly requested another language.

Success Criteria

A good result should:

  • save the user time
  • surface only the highest-value AI developments
  • clearly distinguish original announcements from media reporting
  • avoid duplicate noise
  • provide usable insight for product, strategy, and industry tracking
  • produce a clean, readable Chinese daily brief

Task Scheduling

This skill is designed to be run on a daily schedule. To automate it, set up a cron job that triggers this skill once per day (recommended time: 07:00–09:00 Asia/Shanghai).

Recommended cron configuration:

  • Schedule: cron expression 0 7 * * * with timezone Asia/Shanghai
  • Session target: isolated (do not run in the main conversation session)
  • Delivery: none (results are written directly to Tencent Docs; no Feishu push needed)

Output targets:

  • Daily AI news doc: https://docs.qq.com/aio/DUEtYTUNWaVNxeXN4
  • AI knowledge base: https://docs.qq.com/aio/DUG9ocVBZU01sRmtG

Use the Tencent Docs API (mcporter call "tencent-docs" "smartcanvas.append_insert_smartcanvas_by_markdown") to write results. Do not send Feishu messages from the scheduled session.


Example Invocation

  • Generate today’s AI news daily brief.
  • Create an AI news digest for the last 24 hours.
  • 帮我整理最近24小时的AI新闻日报。
  • 基于既定信源生成今日AI新闻简报。

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

93.49%
按下载量换算2,226

安全审计

VirusTotal

通过

ClawScan

可疑

Static analysis

通过

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills