Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Wow Daily News

v1.0.1

每日日报自动生成。每天 18:00 自动生成包含魔兽世界新闻、NGA 热帖、今日美图的飞书文档,并推送到飞书和微信。

0· 108·1 current·1 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 lorexxar/wow-daily-news.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Wow Daily News" (lorexxar/wow-daily-news) from ClawHub.
Skill page: https://clawhub.ai/lorexxar/wow-daily-news
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
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 wow-daily-news

ClawHub CLI

Package manager switcher

npx clawhub@latest install wow-daily-news
Security Scan
VirusTotalVirusTotal
Pending
View report →
OpenClawOpenClaw
Suspicious
high confidence
!
Purpose & Capability
The SKILL.md describes a multi-step pipeline (collectors, NGA detail scraping, Xiaohongshu beauty fetcher, Feishu/WeChat push). The package only includes one script (scripts/wow_daily_news.py) that implements EXWIND news scraping and local markdown output. The runtime instructions call many other scripts/paths (e.g. ~/.openclaw/workspace/scripts/daily_report_generator.py, ~/.openclaw/workspace/skills/daily-beauty/daily_beauty.py) and expect local services/binaries that are not included. That discrepancy indicates the bundle is incomplete or misdescribed.
!
Instruction Scope
SKILL.md instructs the agent to run many shell commands: curl to localhost:18060 and possibly nohup to start ~/xiaohongshu-mcp, agent-browser to open pages and take snapshots, read/write many files under ~/.openclaw/workspace and /tmp, and send messages to specific Feishu/WeChat targets. These steps reference local binaries/services and data files outside the shipped files. The instructions also require copying image files to /tmp and use platform-specific helper commands (feishu_create_doc, feishu_doc_media, message action=send) — none of which are defined or included. Running the instructed commands could start arbitrary binaries in the user's home or interact with system services.
Install Mechanism
There is no install spec (instruction-only). That lowers the risk of remote code being fetched during install, but the runtime relies on an existing environment with many scripts/binaries and system services. The absence of an install step means the skill assumes pre-existing local components (which are not shipped) — a design mismatch that can cause silent failures or unexpected behavior if the environment contains different binaries.
!
Credentials
The skill declares no required environment variables or credentials, yet the instructions embed hardcoded push targets (Feishu ou_2f7b6746..., a WeChat account ID) and call helper commands that must have credentials configured elsewhere (feishu_create_doc, feishu_doc_media, message action=send). The pipeline also expects writable access to many config/data paths under the user's home and to start binaries (~/xiaohongshu-mcp). Requiring broad filesystem and service access without clear credential configuration is disproportionate and surprising to an installer.
Persistence & Privilege
The skill does not request always:true or other elevated platform privileges. It does instruct starting a user-level binary (nohup ~/xiaohongshu-mcp) and assumes systemd-managed Xvfb, but it does not modify other skills' configs or claim permanent inclusion.
What to consider before installing
Do not install or run this skill unless you understand and control the target environment. Issues to consider before proceeding: 1) Incoherent bundle: the SKILL.md refers to many scripts, data files, and a local MCP service that are not included. The shipped script only fetches EXWIND news — it does not implement the full pipeline. 2) Runtime commands will run shell commands (nohup, curl) that can start arbitrary binaries in ~/xiaohongshu-mcp or other paths; verify those binaries yourself. 3) The skill hardcodes Feishu and WeChat target IDs and relies on helper commands that must have credentials configured elsewhere — confirm these behaviors are acceptable and inspect how feishu_create_doc / message are implemented in your agent runtime. 4) If you want to test, run in an isolated/sandboxed environment and inspect or stub the referenced scripts/services first. Prefer requesting the author to provide the missing scripts, or a clear README explaining required local services and what credentials are needed. If you can't verify the missing components and recipients, treat this skill as risky.

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

latestvk972vc3x48e4yd9ed9g74jsp1h83z2g5
108downloads
0stars
2versions
Updated 4w ago
v1.0.1
MIT-0

每日日报

每天自动生成综合日报,汇总魔兽世界相关资讯,写入飞书文档并推送。

触发词

  • "今日日报"
  • "每日新闻"
  • "生成日报"
  • "日报"

📋 日报内容结构

按顺序包含以下 4 个板块:

  1. 🎮 魔兽世界今日新闻 — EXWIND 蓝帖/热修 + 暴雪中国官网(当日无新闻则跳过)
  2. 📋 艾泽拉斯议事厅 — NGA 热门帖子(3 条,无数据则跳过)
  3. 🏘️ 晴风村 — NGA 晴风村热门帖子(1 条,无数据则跳过)
  4. 🌸 今日美图 — 小红书真人美女博主(9 张图片,获取失败则降级标记)

⏱️ 时间日志(必须执行)

每个 Step 开始和结束时,必须执行以下命令记录时间:

echo "[Step N] start $(date '+%H:%M:%S')"
echo "[Step N] end $(date '+%H:%M:%S')"

这用于排查超时问题,不可跳过


🚨 Agent 执行步骤

Step 1: 运行数据收集脚本

echo "[Step 1] start $(date '+%H:%M:%S')"
python3 ~/.openclaw/workspace/scripts/daily_report_generator.py
echo "[Step 1] end $(date '+%H:%M:%S')"

脚本只负责收集基础数据(标题、链接、URL),不抓详情

Step 2: 解析 JSON 输出

echo "[Step 2] start $(date '+%H:%M:%S')"

从 Step 1 的输出中解析 JSON:

{
  "success": true,
  "date": "2026-03-18",
  "title": "每日日报 - 2026年03月18日",
  "news": [{"title": "...", "url": "...", "date": "...", "type": "...", "source": "..."}],
  "nga_aiz": [{"title": "...", "url": "...", "tid": "...", "replies": 0}],
  "nga_qfc": [{"title": "...", "url": "...", "tid": "...", "replies": 0}],
  "beauty": {"success": true/false, "error": "...", "blogger": {...}, "images": [...]},
  "news_count": 0,
  "nga_count": 4
}
echo "[Step 2] end $(date '+%H:%M:%S')"

Step 3: 处理美图(优先处理,失败可降级)

echo "[Step 3] start $(date '+%H:%M:%S')"
  1. 检查 beauty.success

    • true → 跳到 Step 4
    • false → 进入自动修复流程
  2. 自动修复流程最多重试 1 次):

    a. 检查小红书 MCP 服务是否运行:

    curl -s --connect-timeout 3 http://localhost:18060/mcp
    
    • 无响应 → 启动:
      cd ~/xiaohongshu-mcp && nohup ./xiaohongshu-mcp-linux-amd64 >> mcp.log 2>&1 & sleep 10 && curl -s --connect-timeout 3 http://localhost:18060/mcp
      

    b. 重新运行美图脚本:

    python3 ~/.openclaw/workspace/skills/daily-beauty/daily_beauty.py
    
    • 成功 → 更新 beauty 数据,继续 Step 4
    • 失败 → 降级处理:标记 beauty_failed = true,继续生成日报(美图板块显示降级提示),不要停止执行
  3. ⚠️ 绝对不要因为美图失败而停止日报生成。 美图只是锦上添花,新闻和 NGA 内容更重要。

注:Xvfb 已配置为 systemd 常驻服务(systemctl status xvfb),无需手动启动。

echo "[Step 3] end $(date '+%H:%M:%S')"

Step 4: 抓取 NGA 帖子详情

echo "[Step 4] start $(date '+%H:%M:%S')"

对 nga_aiz + nga_qfc 中的每条帖子,按回复数从高到低依次处理:

  1. 用 agent-browser 打开帖子 url
  2. 查看帖子发表时间/最后回复时间,超过 7 天则跳过,递推到下一个
  3. 在 7 天内 → 提取摘要(100-200字)+ 下载配图到 ~/.openclaw/workspace/img/
  4. 处理完后关闭浏览器

如果列表为空,跳过对应板块

echo "[Step 4] end $(date '+%H:%M:%S')"

Step 5: 获取新闻完整内容(如果有今日新闻)

echo "[Step 5] start $(date '+%H:%M:%S')"

~/.openclaw/workspace/data/wow_news_history.json 的 by_date 字段按 URL 匹配获取 content 字段。

如果 news 为空(今日无新闻),跳过整个板块

echo "[Step 5] end $(date '+%H:%M:%S')"

Step 6: 生成 Lark-flavored Markdown

echo "[Step 6] start $(date '+%H:%M:%S')"

按各板块格式规则生成完整 Markdown 内容(见下方"各板块详细规则")。

echo "[Step 6] end $(date '+%H:%M:%S')"

Step 7: 创建飞书文档

echo "[Step 7] start $(date '+%H:%M:%S')"

使用 feishu_create_doc

  • title:JSON 中的 title 字段
  • markdown:生成的 Markdown
  • 不要在 markdown 开头加一级标题
echo "[Step 7] end $(date '+%H:%M:%S')"

Step 8: 插入美图(仅美图成功时)

echo "[Step 8] start $(date '+%H:%M:%S')"

仅当 beauty_failed 不为 true 时执行:

  • 使用 feishu_doc_media 逐张插入美图到文档末尾
  • action: insert, type: image, doc_id: Step 7 的 doc_id
  • 图片文件必须先复制到 /tmp/ 目录下(feishu_doc_media 只允许 /tmp 路径)
  • 每张图片间隔 2-3 秒,避免 429 限流

如果美图失败,跳过此步骤。

echo "[Step 8] end $(date '+%H:%M:%S')"

Step 9: 推送文档链接(必须执行!)

echo "[Step 9] start $(date '+%H:%M:%S')"

使用 Step 7 获取的文档 URL,必须同时发送到两个渠道:

  • 飞书:message action=send channel=feishu target=user:ou_2f7b674673f4020ca4a64deda675ccc9 message="📰 每日日报已生成\n\n{url}"
  • 微信:message action=send channel=openclaw-weixin target=o9cq80yHzZS7fs4USr2cLP6z53R4@im.wechat accountId=7785501783cf-im-bot message="📰 每日日报已生成\n\n{url}"
echo "[Step 9] end $(date '+%H:%M:%S')"
echo "=== 日报生成完成 ==="

🚨 如果不执行 Step 9 推送链接,日报就等于没生成!这一步绝对不能跳过!


📐 各板块详细规则

🎮 魔兽世界今日新闻

严格规则:

  1. 只包含今日新闻:仅展示当天发布的新闻,无今日新闻则整个板块跳过
  2. 必须包含完整内容:每条新闻必须读取并展示完整正文
  3. 必须包含配图:每条新闻必须抓取原文中的图片
  4. 不降级:不取旧新闻凑数

格式示例:

## 🎮 魔兽世界今日新闻

共 X 条资讯

---

### 📘 蓝帖:职业调整即将到来
📅 2026-03-17 09:56
🔗 [查看原文](https://exwind.net/post/blue/xxx)

<image url="配图URL" width="600" align="center"/>

完整新闻正文内容...

---

📋 艾泽拉斯议事厅

严格规则:

  1. 时间限制:只选择发表于一周内的帖子。打开帖子页面后,查看帖子的发表时间或最后回复时间,超过 7 天则跳过,递推到回复数排序的下一个帖子
  2. 必须包含 100-200 字摘要:打开帖子详情页抓取正文
  3. 有配图则抓取插入
  4. 数量:3 条,按回复数排序,一周内不足 3 条则有多少展示多少
  5. 无数据则跳过板块

时间判断方式:AI 打开帖子页面时,查看帖子的发表时间或最后回复时间,如果超过 7 天则跳过该帖子。

格式示例:

## 📋 艾泽拉斯议事厅

---

### 1. 帖子标题
🔗 [查看帖子](https://nga.178.com/read.php?tid=xxx)  💬 回复:782

<image url="帖子配图URL" width="600" align="center"/>

帖子正文摘要(100-200字)...

---

🏘️ 晴风村

严格规则:

  1. 时间限制:同议事厅,超过 7 天则递推到下一个帖子
  2. 必须包含 100-200 字摘要:同议事厅
  3. 有配图则抓取插入
  4. 数量:1 条,按回复数排序,一周内不足 1 条则跳过板块
  5. 无数据则跳过板块

格式示例:

## 🏘️ 晴风村

---

### 帖子标题
🔗 [查看帖子](https://nga.178.com/read.php?tid=xxx)  💬 回复:7086

<image url="帖子配图URL" width="600" align="center"/>

帖子正文摘要(100-200字)...

---

🌸 今日美图

正常情况

  • 1 位小红书真人美女博主的 9 张全身照
  • 博主信息:昵称、粉丝数、获赞数

降级处理(美图获取失败时):

## 🌸 今日美图

> 今日美图服务暂时不可用,明日恢复 🌟

⚠️ 美图失败不影响日报生成! 日报以新闻和 NGA 内容为核心,美图为加分项。


📁 相关文件

文件说明
~/.openclaw/workspace/scripts/daily_report_generator.py数据收集脚本(纯数据,不抓详情)
~/.openclaw/workspace/scripts/wow_news_monitor.py新闻监控脚本(实时推送用)
~/.openclaw/workspace/data/wow_news_history.json新闻历史数据
~/.openclaw/workspace/data/daily_report_history.json日报历史记录
/tmp/nga_history_multi.jsonNGA 帖子历史数据
~/.openclaw/workspace/skills/daily-beauty/daily_beauty.py美图获取脚本
~/.openclaw/workspace/skills/wow-daily-news/scripts/wow_daily_news.py魔兽新闻详情抓取脚本
~/.openclaw/workspace/img/图片保存目录
~/xiaohongshu-mcp/小红书 MCP 服务目录

🔗 数据源

EXWIND 魔兽数据挖掘

  • URL: https://exwind.net/
  • 历史文件: data/wow_news_history.json
  • 详情获取: 从 history JSON 的 by_date 字段按 URL 匹配 content

暴雪中国官网

NGA 论坛

  • 艾泽拉斯议事厅: fid=7 | 晴风村: fid=-7955747
  • 历史文件: /tmp/nga_history_multi.json
  • 详情获取: agent-browser 抓取摘要和配图

小红书(美图)

  • 脚本: daily_beauty.py
  • MCP 服务: localhost:18060
  • 依赖: Xvfb 虚拟显示 (DISPLAY=:99)

📐 飞书文档格式要求

  • 标题: #####,不重复文档标题
  • 分割线: 板块之间用 ---
  • 加粗: 新闻/帖子标题用 **加粗**
  • 链接: [文字](URL) 语法
  • 图片: <image url="..."/> 语法
  • 图标: 📘蓝帖 🔧热修 📰新闻 📋议事厅 🏘晴风村 🌸美图

⚙️ Cron 配置

  • 任务名: wow-daily-news
  • 调度: cron 0 18 * * *(每天 18:00 北京时间)
  • 超时: 1200 秒(20 分钟)
  • 会话: isolated agentTurn
  • 投递: none(AI 通过 message tool 主动推送)
  • 推送: 飞书 + 微信双渠道

📝 更新记录

日期变更
2026-03-31超时从 600s 增大到 1200s(20分钟)
2026-03-31美图失败降级处理:不再阻塞日报生成,显示"暂时不可用"
2026-03-31每步添加时间日志(echo start/end),方便排查卡点
2026-03-31Xvfb 改为 systemd 常驻服务,移除手动启动逻辑
2026-03-31推送渠道从 LightClaw 改为微信(openclaw-weixin)
2026-03-18美图失败自动修复(启动 MCP),失败则降级处理
2026-03-18脚本重构:只输出基础数据,详情抓取由 AI 完成
2026-03-18新增:NGA 必须摘要+配图
2026-03-18新增:新闻仅今日+完整内容+配图
2026-03-18新增:美图强制板块
2026-03-18支持飞书+LightClaw 双渠道
2026-03-12初始创建

Comments

Loading comments...