Web Learner

自主上网学习技能 - 让 AI 能够主动搜索、浏览和从互联网获取知识。当用户要求了解最新信息、学习新知识、查询新闻、获取某个主题的详细信息,或需要从网络上获取数据时触发此技能。

MIT-0 · Free to use, modify, and redistribute. No attribution required.
0 · 1.4k · 18 current installs · 18 all-time installs
MIT-0
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name and description match the runtime instructions: the SKILL.md explicitly instructs using web_search, web_fetch, and browser tools to perform searches, fetch pages, and render JS-heavy pages. There are no unrelated binaries, credentials, or config paths requested by the skill itself.
Instruction Scope
Instructions stay within the web-retrieval scope (search, fetch, browser). They reference concrete sources (news sites, wttr.in) and require citing sources. A small ambiguity: guidance to "处理反爬虫和登录限制" (handle anti-scraping and login restrictions) is high-level and could be interpreted broadly — the skill does not explicitly instruct bypassing authentication, but this wording could lead an agent to attempt aggressive techniques unless constrained. The SKILL.md also notes web_search requires a Brave API key, but the skill does not declare or request that key.
Install Mechanism
No install spec and no code files — instruction-only. Nothing will be written to disk or fetched by the skill itself.
Credentials
The skill declares no required environment variables or credentials. It does note that some tools (web_search) need a Brave API key and the browser tool needs a running service; those are tool-level requirements rather than skill-level secrets. There are no unexpected secret requests in the skill files.
Persistence & Privilege
always:false and user-invocable:true. The skill does not request permanent presence or attempt to modify agent/system config. Autonomous invocation (disable-model-invocation:false) is the platform default and not a special privilege here.
Assessment
This skill is internally consistent and lightweight (instruction-only). Before installing: (1) confirm how your agent/platform provides the web_search/web_fetch/browser tools and what credentials they require (e.g., Brave API key) — the skill mentions these but doesn't request them; (2) decide whether you want the agent to attempt fetching content behind logins or anti-bot protections and, if not, restrict or clarify that behavior; (3) be mindful of privacy when asking it to fetch URLs or user-specific pages (don’t provide account credentials unless you trust the tool chain); and (4) verify the browser service (if used) runs in a secure, sandboxed environment to avoid exposing local resources.

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

Current versionv1.0.0
Download zip
latestvk97ckgrv1py4htztkdt3chmxe981s2se

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

SKILL.md

Web Learner - 自主上网学习技能

概述

此技能赋予 AI 自主上网搜索和学习的能力,使其能够:

  • 主动从互联网获取最新信息
  • 学习新知识并整合到回答中
  • 查询新闻、天气、事件等信息
  • 获取特定主题的详细资料

触发场景

以下情况应激活此技能:

  1. 用户询问"最新消息"、"新闻"、"今天发生了什么"
  2. 用户要求"了解"、"学习"、"查询"某个主题
  3. 用户要求获取某个链接的详细内容
  4. 需要获取实时信息(天气、股价、赛事等)
  5. 用户要求"汇报"某方面的信息
  6. 需要验证或补充训练数据中没有的知识

搜索策略

1. Web Search (搜索)

使用 web_search 工具进行关键词搜索:

  • 选择精确的搜索关键词
  • 优先选择中文搜索结果(使用 country: "CN"
  • 使用 freshness 参数限定时间范围:pd(今天)、pw(本周)、pm(本月)

2. Web Fetch (抓取)

使用 web_fetch 工具获取页面内容:

  • 提取网页的 markdown 或 text 内容
  • 适用于获取文章、文档、教程等详细内容
  • 注意处理页面加载失败的情况

3. Browser (浏览器)

当以上工具失败时,使用 browser 工具:

  • 需要先启动浏览器服务
  • 可用于处理需要 JavaScript 渲染的页面
  • 支持截图和 UI 交互

学习工作流

1. 理解需求 → 确定需要什么信息
2. 选择工具 → 搜索/抓取/浏览器
3. 获取信息 → 执行查询
4. 整合加工 → 提取关键信息
5. 反馈用户 → 用中文总结呈现

常见场景处理

天气查询

  • 首选 wttr.in 服务(支持城市查询)
  • 备选:weather.com.cn、中国天气网

新闻获取

  • 百度新闻:news.baidu.com
  • 澎湃新闻:www.thepaper.cn
  • 今日头条:www.toutiao.com

获取视频内容

  • 使用 web_fetch 尝试获取视频页面描述
  • 提取视频标题和简介
  • 注意:无法直接播放或理解视频内容

获取文档内容

  • 优先使用 web_fetch 提取可读内容
  • 处理反爬虫和登录限制
  • 提取页面主要文本内容

错误处理

当无法获取信息时:

  1. 尝试备选数据源
  2. 告知用户具体障碍
  3. 提供手动查看的建议链接
  4. 询问是否需要其他帮助

输出格式

向用户呈现信息时:

  • 使用中文
  • 结构清晰,分点陈述
  • 注明信息来源
  • 保持简洁,避免信息过载

工具可用性检查

在执行前检查工具可用性:

  • web_search - 需要 Brave API Key
  • web_fetch - 大部分网页可用
  • browser - 需要浏览器服务运行

如遇工具不可用,尝试备选方案或告知用户。

Files

2 total
Select a file
Select a file to preview.

Comments

Loading comments…