weekly-ai-news

v1.0.0

每周AI前沿动态生成工具。自动抓取 RSS 订阅源的 AI 新闻,筛选应用向内容(剔除技术性太高的内容),生成旧报纸风格的 HTML 简报,支持发送到飞书。使用场景:每周一自动/手动生成上周 AI 应用动态周报,以复古报纸形式呈现并推送到飞书。

0· 141·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for qisee/weekly-ai-news.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "weekly-ai-news" (qisee/weekly-ai-news) from ClawHub.
Skill page: https://clawhub.ai/qisee/weekly-ai-news
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Required binaries: python3, bash
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install weekly-ai-news

ClawHub CLI

Package manager switcher

npx clawhub@latest install weekly-ai-news
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description match the included scripts: RSS fetching (fetch_rss.py), HTML generation (generate_newspaper.py), message formatting (format_feishu_msg.py) and a wrapper script (generate_weekly_report.sh). Required binaries (python3, bash) are reasonable and expected for these tasks. No unexplained env vars, credentials, or config paths are requested.
Instruction Scope
SKILL.md instructs the agent to run only the provided scripts and to optionally invoke openclaw message send; scripts fetch only the declared RSS sources and write local output files. Minor notes: (1) the wrapper and README suggest using openclaw message send but the skill does not include automated Feishu API integration (no tokens requested) — sending is manual/through OpenClaw; (2) date-range logic in fetch_rss.py/get_last_monday is somewhat unusual (may produce unexpected cutoff dates depending on weekday), but this is a correctness issue, not a security one; (3) generated HTML imports Google Fonts and embeds a data: SVG background — these cause harmless external resource loads when opening the HTML.
Install Mechanism
There is no install spec (instruction-only install), and provided code files run with python3/bash on disk. No downloads from third‑party URLs during install are present. Risk from installation is low because nothing fetches arbitrary remote code during install.
Credentials
No environment variables, secrets, or config paths are required. The scripts do not attempt to read unrelated credentials. The only external network activity is fetching the declared RSS feed URLs, which is expected for the stated function.
Persistence & Privilege
The skill is not always-enabled and is user-invocable. It does not modify other skills or global agent configuration. It writes output to a user-specified directory (default ~/weekly-ai-news), which is appropriate for this type of tool.
Assessment
This skill appears to do what it says: fetch configured RSS feeds, filter for application-oriented AI news, produce an HTML 'newspaper', and prepare a Feishu-formatted message. Before installing or scheduling it: (1) review the included RSS source list and remove/add feeds you trust; (2) run the scripts manually once (e.g., generate_weekly_report.sh without --send) to inspect output files (news.json, HTML, message.txt) before any automated sending; (3) note the skill does not automatically send to Feishu because no Feishu credentials are requested — you must configure OpenClaw or provide your own send mechanism; (4) consider running in a limited user account or container if you want isolation; (5) be aware of minor correctness issues (date cutoff logic) and resource loads (Google Fonts when opening HTML). No unrelated credential access or hidden network exfiltration was found.

Like a lobster shell, security has layers — review code before you run it.

Runtime requirements

Binspython3, bash
latestvk97425321h4e73fw6kbhbm1n5183ecz7
141downloads
0stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

每周AI前沿动态

自动抓取、筛选、生成 AI 应用新闻周报,以旧报纸风格呈现,支持发送到飞书。

订阅源

功能

  1. RSS 抓取 - 自动获取订阅源最新内容
  2. 智能筛选 - 识别 AI 相关内容,过滤纯技术文章
  3. 应用导向 - 优先保留产品发布、商业动态、市场趋势等内容
  4. 复古风格 - 生成旧报纸风格的 HTML 简报
  5. 飞书推送 - 生成格式化消息,支持发送到飞书

使用方法

快速生成

weekly-ai-news

生成到默认目录 ~/weekly-ai-news/

指定输出目录

weekly-ai-news /path/to/output

指定天数范围

weekly-ai-news ~/weekly-ai-news 7

生成并发送飞书消息

# 生成后会提示如何发送
weekly-ai-news ~/weekly-ai-news 7 --send

# 或者使用 openclaw 命令直接发送
weekly-ai-news ~/weekly-ai-news 7 && openclaw message send "$(cat ~/weekly-ai-news/message.txt)"

手动分步执行

# 1. 抓取 RSS(过去7天)
python3 ~/.openclaw/skills/weekly-ai-news/scripts/fetch_rss.py --days 7 > news.json

# 2. 生成 HTML
python3 ~/.openclaw/skills/weekly-ai-news/scripts/generate_newspaper.py \
    --input news.json \
    --output weekly-ai-news.html

# 3. 格式化飞书消息
python3 ~/.openclaw/skills/weekly-ai-news/scripts/format_feishu_msg.py \
    --news-json news.json \
    --html-file weekly-ai-news.html > message.txt

筛选逻辑

保留(应用向):

  • 产品发布、上线
  • 融资、投资、并购
  • 商业合作、战略
  • 市场趋势、用户增长
  • 政策法规、监管动态

过滤(技术向):

  • 算法原理、模型训练
  • 神经网络架构
  • 技术实现细节
  • 参数调优、论文解读

输出文件

  • news.json - 原始新闻数据(JSON 格式)
  • weekly-ai-news.html - 旧报纸风格 HTML 简报
  • message.txt - 飞书消息文本(使用 --send 时生成)

旧报纸风格特点

  • 泛黄纸张纹理背景
  • 经典衬线字体(Noto Serif SC)
  • 多栏排版(双栏布局)
  • 首字下沉装饰
  • 头条新闻突出显示
  • 复古色调与边框

飞书消息格式

自动生成的飞书消息包含:

  • 周报标题和期号
  • 头条新闻突出显示
  • 其他新闻列表(最多5条)
  • 来源和日期信息
  • HTML文件路径提示

自动化建议

添加到 crontab 每周一自动执行:

# 每周一上午9点自动生成并发送
0 9 * * 1 /usr/local/bin/weekly-ai-news ~/weekly-ai-news 7 --send

或在 OpenClaw 中设置定时任务:

openclaw cron add --name "weekly-ai-news" \
  --schedule "0 9 * * 1" \
  --command "weekly-ai-news ~/weekly-ai-news 7 --send"

Comments

Loading comments...