Token导航 LogoToken导航TokenDH.com
效率需要联网clawhub未标认证来源可访问clear审计通过

news-brief-automation新闻简报自动化

Agent Skill

news-brief-automation 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 OpenClaw 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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

3,152

周安装

134

GitHub Stars

公开资料未说明

下载量

1,104
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:news-brief-automation(新闻简报自动化)
来源仓库:https://github.com/louishwh/news-brief-automation
安装命令:
openclaw skills install news-brief-automation
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 news-brief-automation

简介

news-brief-automation 用于构建定期新闻摘要工作流程,支持多源聚合。

  • 适用于 Hacker News、GitHub Trending 等平台的自动化简报生成。
  • 通过 clawhub 安装,使用 openclaw skills install news-brief-automation 命令部署。
  • 建议确认调度器配置与输出渠道,注意网络依赖与错误重试机制。
  • 可结合原始 README 了解工作流设计与异常处理逻辑。

SKILL.md

name
news-brief-automation
description
Build, refine, or maintain recurring news-digest workflows that periodically collect items from sources such as Hacker News, GitHub Trending, Hugging Face, RSS feeds, blogs, and specific news sites; deduplicate against the newest prior report; save a Markdown brief in the workspace; and send the same summary to the user. Use when asked to set up scheduled news checks, daily/weekly digests, trend briefings, cron prompt files, or reusable report automation for “check news and send me a summary” tasks.

News Brief Automation

Turn ad-hoc “go check the news and summarize it” work into a repeatable automation made of:

  • one prompt file under automation/cron/
  • one report folder under reports/
  • one stable output format
  • one explicit dedupe rule against the latest saved report

Workflow

  1. Identify the recurring job:

- source set - cadence - selection criteria - output language - save location - delivery target

  1. Prefer one prompt file per job under automation/cron/<job-name>.md.
  2. Create a matching report directory under reports/<topic>/.
  3. Instruct the agent to compare candidates against the newest saved report before drafting the final output.
  4. Require “save first, send second” so future runs always have a concrete baseline for dedupe.
  5. Make the format stable and easy to skim.
  6. If a run has little novelty, emit a short incremental edition instead of padding the report.

What A Good Cron Prompt Must Contain

Each recurring prompt should explicitly define:

  • the goal
  • the schedule context
  • the priority sources to inspect
  • what counts as worth including
  • the required output structure
  • dedupe rules against the newest prior report
  • save path and filename convention
  • send-after-save behavior
  • fallback behavior if a source is unavailable

Use direct instructions. Avoid fuzzy wording like “make it good” or “summarize nicely”.

Good Defaults

Source handling

Prefer primary index pages first, then open individual links only when context is missing.

Common defaults:

  • Hacker News: homepage + newest
  • GitHub: Trending
  • Hugging Face: hot/trending models, datasets, spaces
  • RSS/blogs: feed page first, article page second
  • Generic news sites: homepage/section page first, article page second

Selection logic

Prioritize items that are:

  • newly surfaced
  • rapidly rising
  • widely discussed
  • practically useful
  • unusually opinionated or technically deep

For technical audiences, bias toward:

  • AI
  • developer tools
  • open source
  • infrastructure
  • security
  • data engineering
  • engineering practice

Deduplication

Always inspect the newest report in the target report folder.

Keep only:

  • new items
  • items with clear rank/momentum change
  • items with meaningful new developments
  • previously seen items that deserve a shorter refreshed mention

Drop or compress:

  • unchanged repeats
  • low-signal filler
  • items included yesterday/last run with no meaningful change

If most candidates repeat prior coverage, explicitly publish an incremental short edition.

Output structure

Prefer Markdown with:

  • dated title
  • short “what matters now” section
  • numbered list of items
  • link for each item
  • one-line reason to care
  • optional discussion angle / watchpoint

Keep tone concise and opinionated enough to be useful.

File And Naming Conventions

Use this layout unless the user asks otherwise:

automation/cron/<job-name>.md
reports/<topic>/YYYY-MM-DD.md
reports/<topic>/YYYY-MM-DD-HH00.md

Use date-only filenames for once-per-day reports. Use hour-stamped filenames for multiple runs per day.

Common Patterns

Daily digest

Use when the user wants one higher-signal summary per day.

Recommended traits:

  • 5-10 items
  • more synthesis
  • a “today’s observation” section
  • date-only filenames

Intraday brief

Use when the user wants multiple updates per day.

Recommended traits:

  • 3-8 items
  • stronger dedupe
  • incremental mode when nothing major changed
  • hour-stamped filenames

Mixed-source trend watch

Use when combining 2-4 complementary sources like GitHub + Hugging Face + HN.

Recommended traits:

  • sectioned output per source
  • final cross-source observations
  • avoid repeating the same project across sections unless the angle differs

Editing Existing Automations

When converting an existing one-off prompt into a reusable workflow:

  1. Preserve the parts that already work:

- source order - output shape - report path

  1. Remove accidental details that do not generalize.
  2. Keep the reusable operating logic:

- fetch - select - dedupe - save - send

  1. Move concrete examples and templates into references/ so SKILL.md stays lean.

Quality Bar Before Packaging Or Publishing

Before packaging or publishing:

  • keep SKILL.md concise and trigger-focused
  • ensure the description clearly says when the skill should activate
  • provide at least one reusable prompt template in references/
  • avoid extra docs like README or changelog files inside the skill
  • package the exact skill folder
  • reject vague prompts that do not specify dedupe or save paths

References

  • Read references/examples.md for concrete example automations.
  • Read references/templates.md for reusable cron prompt templates.

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

96.7%
按下载量换算1,068

安全审计

VirusTotal

未展示

ClawScan

通过

Static analysis

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。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