福利吧

获取福利吧 (https://www.wnflb2023.com/) 指定板块的帖子列表（标题、内容、链接），进入帖子提取正文内容，最后由 AI 归纳总结。

Security Scan

The skill generally matches its stated purpose (scraping a forum) but asks for and manipulates a local forum cookie file and executes untrusted JavaScript in a Node VM — reasonable for scraping but raises privacy and execution-sandboxing concerns that you should review before installing.

Purpose & Capability

Name, description, SKILL.md, package.json, and code files all consistently implement a forum scraper for https://www.wnflb2023.com. Dependencies (cheerio, undici) and the npm install instruction are appropriate for HTML fetching/parsing.

Instruction Scope

SKILL.md explicitly requires reading a local cookie.txt and running node main.js to scrape and check-in. That is within scope for a logged-in forum scraper, but the skill instructs the agent to read/write a cookie file in the skill directory (session tokens) and to extract potentially sensitive content (links/passwords). The README's demand that cookie.txt be "only readable by this skill" is odd and not enforceable by the skill itself.

Install Mechanism

There is no formal install spec; SKILL.md requires running 'npm install' which will fetch packages from the npm registry (cheerio, undici). This is a normal, moderate-risk action — no third-party download URLs or extract steps are present.

Credentials

The skill does not request environment variables but requires a local cookie file containing forum session cookies. Cookies are effectively credentials and will be read and updated by the skill (readCookie/writeCookie). This is proportionate for a logged-in scraper, but users should be aware that these cookies may grant access to their forum account and should not contain other site/session tokens.

Persistence & Privilege

The skill does not request 'always: true' or any elevated platform privileges. It reads/writes its own cookie file in its directory and does not modify other skills or global agent settings.

What to consider before installing

This skill is a coherent forum scraper, but it needs a local cookie file (cookie.txt) containing your forum session cookies — treat that like a password. Before installing: (1) verify you trust the skill source and the forum; (2) consider creating a throwaway forum account or limited-session cookie instead of using your main account; (3) run the skill in an isolated environment/container; (4) inspect cookie.txt contents and file permissions (avoid storing other credentials); (5) be aware the code executes untrusted page JavaScript inside Node's vm to resolve redirects — the sandbox reduces but may not eliminate risk. If you cannot accept the skill having read/write access to a forum session cookie, do not install.

安装说明

• 操作系统不限
• 系统需要有node环境，版本需要22及以上. 可以在环境变量中检测是否已经安装了node
• 需要执行npm install来安装依赖的包

使用手册

何时使用本 Skill

• 用户说：“帮我看一下发现之门/综合讨论/购物网赚的最新帖子”
• 用户说："福利吧最新帖"
• 用户说："福利吧签到"

核心工作流（4步循环 + AI 总结）

1: 检查cookie.txt文件 重要：以下操作都在本Skill目录 ./ 下进行（即包含此SKILL.md文件的目录）。 - 使用 read_file 工具读取 ./cookie.txt 文件

• 检查文件是否存在且内容非空
• 必须确保cookie.txt文件仅对本skill可读。

2. 确定论坛板块和抓取数量

   • 板块fid列表:
     • 发现之门:fid=2
     • 综合讨论:fid=40
     • 购物网赚:fid=37
     • 网游分享区:fid=67
     • 交易平台:fid=44
     • 福禄娃之家:fid=39
   • 如果用户说“看下福利吧最新帖子”，默认使用发现之门的fid
   • 如果用户说“看下福利吧综合讨论的帖子”则fid为40
   • 抓取数量(count)默认为10条，如果用户有指定则使用用户指定的数量

3. 获取帖子及内容

   • 执行以下代码获取输出JSON

   ⚠️ 执行注意事项

   脚本调用时禁止使用链式运算符（&&、;、|| 等） 错误的例子： cd ~/.openclaw/workspace/skills/wnflb && node main.js 正确的列子： node ~/.openclaw/workspace/skills/wnflb/main.js 即：始终使用绝对路径调用 node + 脚本，不需要 cd && 前缀。

     node ~/.openclaw/workspace/skills/wnflb/main.js scrape --fid fid --count count

 - fid为对应板块
 - count为抓取的帖子数量

• 程序会返回json的格式内容

{
  "forumTitle": "发现之门 -",
  "forumUrl": "https://www.wnflb2023.com/forum.php?mod=forumdisplay&fid=2&mobile=2",
  "fid": 2,
  "count": 1,
  "items": [
    {
      "title": "有多少人还没有副业？",
      "url": "https://www.wnflb2023.com/forum.php?mod=viewthread&tid=274938&extra=page%3D1&mobile=2&_dsign=7950febd",
      "content": "副业已经伴随我快10年了，中途也换过无数个副业了，总感觉没有副业心理不踏实\n\n不做副业的是不是主业收入相当可观了"
    }]
}

4. AI内容摘要并输出

• 将content内容传入LLM（你就是 AI）
• |内容摘要|：200字内,去掉html的换行等标签，总结要突出核心福利和有效信息,有网盘的链接地址和密码需要提取出来并显示,如果有百家姓等内容也需要把内容显示出来,如果有多个网盘链接，列出每个链接及对应密码。
• 严格按照以下结构输出Markdown的表格格式

### <板块名称>最新内容
| 序号 | 标题 | 内容摘要 |
|------|------|----------|
| [1](<url>) | xxxxxxx| xxxxxx| 

论坛签到

• 论坛签到是独立的一个功能,同样的需要先满足准备Cookie的工作（参见上面的流程1）
• 执行脚本获取结果即可

node ~/.openclaw/workspace/skills/wnflb/main.js checkin

返回json。例如

{
  "ok": true,
  "entryUrl": "https://www.wnflb2023.com/misc.php?mod=mobile",
  "checkinUrl": "https://www.wnflb2023.com/plugin.php?id=fx_checkin:checkin&formhash=1cbf33ef&inajax=1",
  "message": "已签到"
}