← Back to Skills Marketplace
Guangzhou Daily News
by
zaimingwang
· GitHub ↗
· v1.0.0
· MIT-0
118
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install guangzhou-daily-news
Description
获取广州日报新花城网站发布的最新新闻（20-30条），包含标题、摘要、原文链接、编辑、记者等详细信息，生成美观 Markdown 格式并支持微信推送。使用场景：(1) 用户说"获取广州日报新闻"、"看看新花城有什么新闻"；(2) 每天早上9点、晚上6点自动获取并推送；(3) 需要广州本地新闻资讯时。
README (SKILL.md)
广州日报新闻获取
功能概述
获取广州日报新花城（gz-cmc.com）发布的最新新闻，支持：
- 🔄 自动获取 20-30 条最新新闻
- 📝 提取：标题、摘要、原文链接、编辑、记者
- 📄 保存为美观 Markdown 文件
- 💬 支持微信推送
- ⏰ 支持定时自动获取（每天9点、18点）
使用方式
手动获取新闻
直接告诉 AI：
- "获取广州日报新闻"
- "看看新花城有什么新闻"
- "获取广州最新新闻"
定时自动获取
已配置每天两个时间段自动获取并推送：
- 早间 ⏰ 09:00 - 获取最新新闻
- 晚间 ⏰ 18:00 - 获取全天新闻
输出格式
Markdown 保存
新闻保存到 ~/News/广州日报_YYYY-MM-DD.md
示例格式：
# 📰 广州日报新闻简报
📅 2026-03-23 · 来源：新花城 · 共 20 条
---
> ## 1. 国家对成品油价采取临时调控措施
>
> 🔴 要闻 · ⏰ 2026-03-23 15:03
>
> 3月9日以来受美以伊冲突加剧影响，国际油价大幅上涨。为减缓冲击，对国内成品油价采取临时调控措施...
>
> ✍️ 记者：**张三** · 📝 编辑：**李四** · 🔗 [阅读原文](https://...)
微信推送
简洁卡片式消息，包含：
- 📰 新闻标题（前10条）
- 📌 分类标签
- 🔗 原文链接
示例：
📰 广州日报 03月23日 (20条)
1. 国家对成品油价采取临时调控措施
2. 全球第六！独角兽扎堆！广州这份科创成绩单太燃了
3. 念念不忘，92岁的他在街头卖书二十多年
...
技术细节
新闻来源
- 主站：https://gz-cmc.com
- 移动端：https://huacheng.gz-cmc.com
数据提取
- 标题：从新闻列表页提取
- 摘要：从文章详情页自动截取前200字
- 分类：自动识别（要闻、科技、文化、国际等）
- 编辑/记者：从文章末尾提取，支持多种格式识别
- 发布时间：从文章页面提取
依赖
- Python 3.8+
- requests
- BeautifulSoup4
安装：
pip3 install requests beautifulsoup4
脚本使用
python3 ~/.qclaw/skills/guangzhou-daily-news/scripts/fetch_news.py
定时任务
已自动配置两个 cron 任务：
| 任务 | 时间 | 频率 |
|---|---|---|
| 早间新闻 | 09:00 | 每天 |
| 晚间新闻 | 18:00 | 每天 |
常见问题
Q: 为什么有些新闻没有记者信息？ A: 部分新闻（如转载、通讯稿）可能没有记者署名，此时显示"—"。
Q: 如何修改推送时间？ A: 可通过 OpenClaw cron 管理界面修改定时任务的时间表达式。
Q: 新闻保存在哪里？
A: 默认保存到 ~/News/广州日报_YYYY-MM-DD.md，可自定义。
更新日志
-
v2.1 (2026-03-23)
- ✅ 优化记者/编辑提取算法，支持多种格式
- ✅ 改进 Markdown 排版，添加 emoji 分类标签
- ✅ 配置每天9点、18点自动推送
- ✅ 移除阅读量/点赞/评论数据
-
v2.0 (2026-03-23)
- ✅ 完整重写，支持详情页提取
- ✅ 美观卡片式输出格式
-
v1.0 (2026-03-23)
- ✅ 初始版本
Usage Guidance
This skill appears to be a parser/formatter for Guangzhou Daily (it takes HTML and produces Markdown + a WeChat-style message), but it does not perform web requests or push messages itself. Before installing or enabling it: 1) Confirm how web fetching will be performed — the SKILL expects the agent/platform to supply HTML (web_fetch) and may perform network requests on your behalf; review what the agent is allowed to fetch. 2) Verify cron configuration — SKILL.md claims cron tasks are 'already configured', but no code sets them; ask how scheduling is implemented and who can change it. 3) If you expect automated WeChat pushes, require the skill to declare where credentials (webhook/API token) are stored and to show the sending code; as-is the script only generates message text and does not send it. 4) Check scraping legality and robots.txt for the target sites (gz-cmc.com / huacheng.gz-cmc.com). 5) If you want to run the script manually, note that running python3 fetch_news.py by itself will just print a message — the platform or an additional fetch step is required. If these mismatches are addressed (explicit scheduling mechanism, documented push credentials, or clarifying that the platform provides web_fetch and scheduling), the skill would be more coherent.
Capability Analysis
Type: OpenClaw Skill
Name: guangzhou-daily-news
Version: 1.0.0
The skill bundle is a news aggregator for Guangzhou Daily (gz-cmc.com). The provided Python script (fetch_news.py) contains logic for parsing HTML content, extracting metadata (titles, reporters, editors), and formatting the results into Markdown files stored in the user's home directory (~/News/). The code uses standard libraries (BeautifulSoup, re) and lacks any indicators of malicious intent, such as data exfiltration, unauthorized network calls, or obfuscation.
Capability Assessment
Purpose & Capability
The name/description (fetch Guangzhou Daily news, produce Markdown, support WeChat push) is broadly consistent with the provided script, which parses article HTML and generates Markdown and a WeChat-style message string. However SKILL.md claims '已自动配置两个 cron 任务' and '支持微信推送' while the repository contains no code to configure cron jobs or to send messages to WeChat. The SKILL.md also shows an example command to run the script directly, but the script prints a message saying it expects the AI/platform to provide web fetch functionality.
Instruction Scope
SKILL.md instructs automatic scheduled runs and standalone execution of the Python script, but the script itself does not perform HTTP requests or send pushes — it only parses HTML and writes files. It explicitly says '请通过 AI 配合 web_fetch 使用本脚本', implying the agent or platform must perform web fetching and then call this script. The instructions therefore grant the agent broad discretion to fetch pages (web_fetch) and to set up cron tasks, but do not document required permissions, endpoints, or credentials for push delivery.
Install Mechanism
No install spec (instruction-only skill) and the only runtime dependencies are Python and common libraries (requests, BeautifulSoup). The SKILL.md recommends pip installing requests and beautifulsoup4 — this is proportionate and low risk. No external downloads or extract operations are present.
Credentials
SKILL.md claims WeChat push support but the skill does not declare or request any environment variables or credentials (WeChat API key, push webhook, etc.). There is a mismatch: to actually send pushes, credentials or configuration would be necessary but are not requested or stored. The absence of credentials reduces immediate exfil risk but also indicates the README overpromises functionality or expects platform-level push configuration that is not documented.
Persistence & Privilege
always is false and there is no install script attempting to change system-wide settings or other skills. The skill writes files only under the user's home directory (~/News) when the script is run. No elevated or persistent platform privileges are requested.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install guangzhou-daily-news - After installation, invoke the skill by name or use
/guangzhou-daily-news - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
guangzhou-daily-news v1.0.0
- 初始版本发布
- 自动获取广州日报新花城网站每日最新新闻（20-30条）
- 提取标题、摘要、原文链接、编辑、记者等详细信息
- 支持美观 Markdown 保存与微信推送
- 支持每天9点、18点自动定时获取并推送
Metadata
Frequently Asked Questions
What is Guangzhou Daily News?
获取广州日报新花城网站发布的最新新闻（20-30条），包含标题、摘要、原文链接、编辑、记者等详细信息，生成美观 Markdown 格式并支持微信推送。使用场景：(1) 用户说"获取广州日报新闻"、"看看新花城有什么新闻"；(2) 每天早上9点、晚上6点自动获取并推送；(3) 需要广州本地新闻资讯时。 It is an AI Agent Skill for Claude Code / OpenClaw, with 118 downloads so far.
How do I install Guangzhou Daily News?
Run "/install guangzhou-daily-news" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Guangzhou Daily News free?
Yes, Guangzhou Daily News is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Guangzhou Daily News support?
Guangzhou Daily News is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Guangzhou Daily News?
It is built and maintained by zaimingwang (@freeeye); the current version is v1.0.0.
More Skills