Token导航 LogoToken导航TokenDH.com
研究检索敏感数据clawhub未标认证来源可访问clear审计提醒

ai-high-quality-briefingAI 高品质简报

Agent Skill

ai-high-quality-briefing 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

630

周安装

26

GitHub Stars

公开资料未说明

下载量

206
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-high-quality-briefing(AI 高品质简报)
来源仓库:https://github.com/kelcey2023/ai-high-quality-briefing
安装命令:
openclaw skills install ai-high-quality-briefing
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-high-quality-briefing

简介

用于查找、检索和筛选相关信息,适合根据关键词、任务场景或来源线索快速定位候选结果。

  • 为人工智能、技术和全球扩张主题生成高质量的中文信息简报。
  • 通过 clawhub 安装,使用 openclaw skills install ai-high-quality-briefing 命令部署。
  • 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。
  • 适用宿主包括 OpenClaw,接入前应确认版本、权限和运行环境要求。

SKILL.md

name
ai-high-quality-briefing
description
Generate high-quality Chinese information briefings for AI, technology, and global expansion topics. Use when the user wants daily news digests, weekly deep briefings, trend tracking, high-signal information filtering, publishable newsletter copy, WeChat/public-account style summaries, internal briefing notes, or recurring scheduled brief generation. Especially use for themes like AI updates, frontier models, arXiv papers, GitHub projects, Hacker News discussions, Hugging Face releases, tech media trends, overseas market intelligence, and 'search the whole web first, then filter for high quality'.

AI / 科技 / 出海高质量信息简报

Produce concise, high-signal Chinese briefings from broad web research.

Read /Users/youke/.openclaw/workspace/skills/ai-high-quality-briefing/references-query-patterns.md when you need search starters. Read /Users/youke/.openclaw/workspace/skills/ai-high-quality-briefing/references-scoring-and-format.md when you need ranking, de-duplication, or publishable copy structure. Read /Users/youke/.openclaw/workspace/skills/ai-high-quality-briefing/references-no-key-sources.md when web_search is unavailable or the environment has no search API key.

Core workflow

  1. Clarify the briefing mode if the user did not specify it:

- 每日快讯: 5–12 items, short and fast - 每周深度版: 3–6 themes, grouped analysis, stronger commentary

  1. Define the scope:

- Topic: AI / technology / overseas expansion, or a narrower subtopic the user gives - Time window: today / last 24h / this week / custom range - Audience: public post / internal briefing / founder / investor / operator

  1. Discover candidate items using one of these paths:

- Preferred path: web_search for broad discovery, then web_fetch for verification - No-key fallback path: if web_search is unavailable, use curated public pages and web_fetch directly; optionally use browser snapshot on already-open or login-required pages

  1. Score and rank candidate items before writing.
  2. Remove duplicates, weak reposts, and thin summaries.
  3. Write in Chinese with an editor’s judgment, not a raw link dump.
  4. If the user wants publishable copy, produce platform-ready versions instead of one neutral block.

No-key fallback policy

If web_search fails because the environment has no API key:

  1. Do not stop.
  2. Switch to the fallback source list in references-no-key-sources.md.
  3. Fetch a small set of strong public pages directly with web_fetch.
  4. Prefer official or source-near pages over commentary pages.
  5. Be explicit in the output that this run is based on curated public sources instead of broad web search, if that limitation materially affects coverage.

Source priorities

Prefer this order unless the user says otherwise:

  1. arXiv
  2. GitHub
  3. Hacker News
  4. Hugging Face
  5. Mainstream tech media
  6. Official product / company blogs

When covering overseas expansion, also prioritize:

  • official company announcements
  • product pages / docs
  • reputable global business or tech reporting
  • region-specific market signals if relevant

Output standards

Always include:

  • a sharp title
  • a short opening summary
  • the main items grouped logically
  • why each item matters
  • source links

Do not just summarize. Add editorial value:

  • what changed
  • why it matters
  • who should care
  • whether it is signal or noise

Default output templates

Template A — 每日快讯

Use this for fast-moving updates.

Structure:

# 今日 AI / 科技 / 出海简报|YYYY-MM-DD

一句话总览:用 1–2 句话总结今天最值得关注的主线。

## 1. 最重要的 3 条
### [标题]
- 发生了什么:
- 为什么重要:
- 适合谁关注:
- 来源:

## 2. 值得关注的新工具 / 新模型 / 新项目
- [名称]:一句话说明 + 为什么值得看
- 来源:

## 3. 可直接拿去发的短文案
### 朋友圈/社媒版
...

### 公众号/内部群版
...

Template B — 每周深度版

Use this for synthesis and stronger interpretation.

Structure:

# 本周 AI / 科技 / 出海高质量简报

## 本周结论
- 结论 1
- 结论 2
- 结论 3

## 一、本周最重要趋势
### 趋势标题
- 发生了什么
- 背后逻辑
- 影响判断
- 代表案例
- 来源

## 二、值得重点追踪的项目 / 公司 / 论文 / 工具
### 名称
- 核心信息
- 亮点
- 局限或风险
- 适用场景
- 来源

## 三、对内容运营 / 产品 / 出海的启发
- 启发 1
- 启发 2
- 启发 3

## 四、可直接发布的文案
### 公众号版
...

### 朋友圈版
...

### 内部汇报版
...

Writing style

Use:

  • Chinese, clean and publishable
  • Dense but readable language
  • Concrete judgment instead of hype
  • Short paragraphs and bullets

Avoid:

  • empty adjectives like “震撼”“炸裂” unless the user explicitly wants hype
  • too many items with no hierarchy
  • long untranslated English passages
  • fake certainty when evidence is weak

Ranking heuristic

When you have many candidates, rank by:

  1. importance
  2. novelty
  3. practical relevance
  4. discussion / traction quality
  5. credibility of source

A smaller high-quality briefing beats a larger noisy one.

Publishable copy generation

When the user wants “可发布文案”, provide at least 2 versions:

  • a shorter social version
  • a fuller article/internal briefing version

If the user’s platform is unknown, default to:

  • 社媒短版: 80–180 Chinese characters
  • 长版摘要: 300–800 Chinese characters

Scheduling guidance

When the user wants recurring briefings:

  • use cron
  • prefer reminders that clearly state the theme, cadence, and desired format
  • mention whether it is a daily fast digest or weekly deep digest
  • include the latest scope in the reminder text so future runs stay consistent

Example reminder text:

  • 提醒:生成今日 AI / 科技 / 出海高质量信息简报。先全网搜索,再筛高质量信号,优先 arXiv、GitHub、Hacker News、Hugging Face 和主流科技媒体;若搜索工具不可用,则改用公开页面与站点直抓;输出中文每日快讯,并附可直接发布的短版与长版文案。
  • 提醒:生成本周 AI / 科技 / 出海高质量信息深度简报。按趋势分组,突出最重要变化、项目、论文、工具与出海启发;若搜索工具不可用,则改用公开页面与站点直抓;输出中文每周深度版,并附公众号版、朋友圈版、内部汇报版文案。

Tool use pattern

Path A — With search key

  1. Start with web_search using focused topic queries
  2. Collect 5–15 promising candidates
  3. Use web_fetch to verify the best ones
  4. Write only after filtering

Path B — No search key

  1. Open the fallback source list
  2. Fetch 1–3 public pages per source with web_fetch
  3. Extract only fresh, concrete items
  4. Merge duplicates
  5. Write only after filtering

Suggested query patterns:

  • latest AI model releases this week
  • best new GitHub AI tools this week
  • important arXiv AI papers this week
  • Hacker News AI startup tools this week
  • global expansion SaaS AI trends this week

When a topic is broad, split queries by source or theme instead of relying on one giant search.

Quality bar before finalizing

Before sending, check:

  • Is each item actually worth the reader’s time?
  • Did I explain why it matters?
  • Did I avoid repeating the same story from multiple angles?
  • Is the output already usable as a post, summary, or internal brief?

If not, tighten it.

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

91.17%
按下载量换算188

安全审计

VirusTotal

通过

ClawScan

可疑

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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