← 返回 Skills 市场
sonpham1909

Google Research Pro

作者 sonpham1909 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
113
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install google-research-pro
功能描述
Performs advanced Google searches with Playwright for OpenClaw technical resources and summarizes results for Telegram channel delivery in Markdown format.
使用说明 (SKILL.md)

Google_Research_Pro\r

\r Performs advanced Google searches using Playwright to bypass bot detection, focusing on technical documentation, GitHub repositories, and usage guides for OpenClaw projects. Results are summarized and sent to a specified Telegram channel.\r \r

Overview\r

\r

  • Skill name: Google_Research_Pro\r
  • Purpose: Enhance web searching with robust, bot-bypass-capable Playwright automation to fetch technical OpenClaw resources from Google.\r
  • Output: A Markdown-formatted summary of top results, delivered to Telegram channel (via a separate workflow step).\r \r

Workflow\r

\r This skill implements a workflow to enhance web searching capabilities by simulating real user behavior with Playwright.\r \r

1. UserInput\r

\r

  • Description: Receives the search keyword from the user.\r
  • Input: keyword (string)\r \r

2. ExecutePythonScript\r

\r

  • Description: Executes a Python script that uses Playwright to perform an advanced Google search for the given keyword, targeting technical OpenClaw resources.\r
  • Input: keyword (string) from UserInput\r
  • Action: Run Python script\r
    • Command: python C:\Users\Admin\OneDrive\Desktop\LearnOpCL\bot.py "{keyword}"\r
    • Working directory: C:\Users\Admin\OneDrive\Desktop\LearnOpCL\r
    • Notes: Ensure Playwright browsers are installed (playwright install)\r \r
  • Output: search_results (structured data, e.g., JSON) containing titles and summaries from the top 3-5 results\r \r

3. ProcessResults\r

\r

  • Description: Formats the extracted search results into a Markdown string.\r
  • Input: search_results\r
  • Action: Transform into a concise Markdown summary\r \r
  • Output: formatted_summary (Markdown string)\r \r

4. SendToTelegram\r

\r

  • Description: Sends the formatted summary to the specified Telegram channel\r
  • Input: formatted_summary\r
  • Action: Send message via Telegram bot/channel\r
    • Destination: Telegram channel/user ID (example: 6830424983)\r
    • Content: formatted_summary (Markdown)\r \r
  • Output: Confirmation of sending or error status\r \r

Implementation Notes\r

\r

  • ExecutePythonScript assumes bot.py can accept a keyword argument and return a structured result.\r
  • The top-3-5 results should include: title, URL, and a short summary/snippet.\r
  • The formatting should be Markdown-friendly for Telegram delivery.\r
  • If the Telegram integration requires a token or bot setup, separate configuration steps are needed.\r \r

Security & Compliance\r

\r

  • This workflow uses Playwright to simulate real user behavior. Ensure compliance with Google terms of service and local laws.\r
  • Do not expose credentials in this file. Use environment variables or a secure vault for sensitive data (e.g., Telegram bot token, Google account credentials if needed).
安全使用建议
Do not install or run this skill until you verify the missing pieces. Specifically: (1) ask the publisher for the bot.py source (the skill currently points to C:\Users\Admin\OneDrive\Desktop\LearnOpCL\bot.py but does not include it); review that code line-by-line before executing it. (2) Require the skill to declare any required environment variables (e.g., TELEGRAM_BOT_TOKEN, TELEGRAM_CHAT_ID) in metadata and confirm where they will be stored. (3) Understand Playwright/browser installation needs and run them in a sandboxed environment; avoid running scripts that claim to 'bypass bot detection' because that may violate Google terms and could be used for abusive scraping. (4) If you must test, run the referenced bot.py in an isolated VM/container with no access to your real secrets or personal files, and monitor network activity. If the author cannot provide the missing code and a clear explanation for why credentials aren't declared, do not trust this skill.
功能分析
Type: OpenClaw Skill Name: google-research-pro Version: 1.0.0 The skill contains a critical shell injection vulnerability in SKILL.md, where the user-provided '{keyword}' is interpolated directly into a system command. It also hardcodes a specific Telegram recipient ID (6830424983) for data delivery and relies on an external Python script (bot.py) located at a specific local path (C:\Users\Admin\OneDrive\Desktop\LearnOpCL) that is not included in the bundle, which is highly irregular for a portable skill.
能力评估
Purpose & Capability
The declared purpose (Playwright-driven Google searches + Telegram summaries) is plausible, but the skill metadata lists no code, no install steps, and no required environment variables. The SKILL.md expects a local Python script at C:\Users\Admin\OneDrive\Desktop\LearnOpCL\bot.py which is not provided by the skill bundle — that is incoherent with an instruction-only skill published with no code.
Instruction Scope
Runtime instructions explicitly tell the agent to execute a local Python script using an absolute path and working directory (OneDrive path). That script would run with the agent's privileges, can access local files and network, and is not included in the skill. The instructions also describe using Playwright to 'bypass bot detection' (techniques that could violate service terms) and to send messages to a Telegram channel, but do not declare or require the Telegram token/chat ID in the metadata.
Install Mechanism
No install spec is provided (instruction-only), which is low-risk in itself. However SKILL.md requires Playwright and its browser installs (playwright install) and a local bot.py — the absence of an install section means the skill assumes local setup that isn't managed or described by the skill bundle.
Credentials
The skill will need sensitive credentials (e.g., Telegram bot token and channel ID) and possibly Google account credentials or proxies, but requires.env is empty and no primary credential is declared. This mismatch means the skill asks the agent to handle secrets without declaring them, and instructs executing an external script that could itself access any environment variables or secrets on the host.
Persistence & Privilege
The skill does not request persistent always:true, which is good, but its runtime instructions require executing arbitrary local code (bot.py) from the user's OneDrive path. That grants the executed code full runtime privileges and potential access to user files and networks — a high-privilege action that is not reflected in the skill metadata.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install google-research-pro
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /google-research-pro 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release of Google_Research_Pro—advanced Google search automation for OpenClaw resources with Telegram delivery. - Automates technical Google searches using Playwright to bypass detection. - Targets documentation, GitHub repos, and usage guides relevant to OpenClaw. - Summarizes top results in Markdown format. - Delivers formatted summaries to a specified Telegram channel. - Designed for flexible integration and secure credential handling.
元数据
Slug google-research-pro
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Google Research Pro 是什么?

Performs advanced Google searches with Playwright for OpenClaw technical resources and summarizes results for Telegram channel delivery in Markdown format. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 113 次。

如何安装 Google Research Pro?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install google-research-pro」即可一键安装,无需额外配置。

Google Research Pro 是免费的吗?

是的,Google Research Pro 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Google Research Pro 支持哪些平台?

Google Research Pro 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Google Research Pro?

由 sonpham1909(@sonpham1909)开发并维护,当前版本 v1.0.0。

💬 留言讨论