← 返回 Skills 市场
wenbing3049

Mirror Website

作者 wenbing3049 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
128
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install mirror-website
功能描述
镜像网站到本地的完整工作流。当用户说镜像网站、下载网站、抓取网站、本地化网站,或者提供一个网址和保存路径时,必须使用此 skill。输入格式为:镜像网站 www.example.com 保存地址 /a/b/c。自动处理代理、递归下载、文件名清理等全流程。即使用户只是提到想把某个网站保存到本地、离线浏览、克隆站点,...
安全使用建议
This skill appears to do what it says (mirror websites) but exercise caution before installing or running it: - Verify the agent environment has the required CLI tools (wget, curl, grep, sed, find, mv/rename, bash, and a Python or other static web server for preview). The SKILL.md relies on these but the skill metadata does not declare them. - The skill explicitly disables robots.txt and skips SSL certificate checks (--no-check-certificate, -e robots=off). That can be legally or ethically problematic and may expose you to unsafe content; only proceed if you understand and accept the implications for the target site. - The instructions create and execute shell scripts in the chosen save path. Review any generated scripts (fetch-query-assets.sh, fix-filenames.sh, etc.) before executing to ensure they do only what you expect and are not pointed at sensitive system paths. - The skill may set http(s)_proxy to a local proxy at 127.0.0.1:7890 if direct access fails. Ensure you trust any local proxy and understand that traffic may be routed through it. - Avoid saving mirrors to sensitive directories (e.g., system paths, home root) and prefer running in a disposable or restricted environment. Consider running the workflow in a sandbox or VM if you are unsure. If you want to proceed, ask the skill to report the exact commands and to show the full contents of any scripts it will write before it executes them; do not allow blind execution of created files.
功能分析
Type: OpenClaw Skill Name: mirror-website Version: 1.0.0 The skill performs high-risk operations including generating and executing shell scripts (fetch-query-assets.sh and fix-filenames.sh) and opening a local network port via a background Python HTTP server. While these actions align with the stated goal of mirroring and previewing websites, the automated execution of agent-created scripts and the hardcoded attempt to use a local proxy (127.0.0.1:7890) represent a significant security risk and potential for unintended side effects on the host system.
能力评估
Purpose & Capability
The SKILL.md actions (curl/wget/creating scripts, cleaning filenames, starting a local HTTP server) are coherent with a website-mirroring purpose. However the skill does not declare required system binaries that it clearly relies on (wget, curl, grep, sed, find, mv/rename, bash, python or similar for preview), which is an omission and reduces transparency.
Instruction Scope
Instructions tell the agent to create and execute shell scripts in arbitrary user-specified directories, scan and rewrite downloaded HTML/CSS/JS, and forcibly ignore robots.txt and SSL certificate validation. While these are related to mirroring, they broaden the agent's actions (writing/executing files, changing network behavior) and include steps (robots=off, --no-check-certificate) that are ethically and security-sensitive.
Install Mechanism
This is instruction-only with no install spec or external downloads, which reduces supply-chain risk. There is nothing being fetched or installed by the skill itself.
Credentials
The skill declares no environment variables or credentials, yet the runtime instructions set http(s)_proxy and rely on local proxy 127.0.0.1:7890 if available. It also implicitly requires many CLI tools that are not listed. The discrepancy between declared requirements (none) and actual required system tools/environment is a proportionality and transparency problem.
Persistence & Privilege
The skill does not request always:true, does not modify other skills, and is user-invocable only. It will write files to user-specified locations, which is expected for this functionality but is not the same as requesting persistent elevated platform privileges.
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 mirror-website
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /mirror-website 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 - 镜像网站到本地的完整工作流,支持离线浏览
元数据
Slug mirror-website
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

Mirror Website 是什么?

镜像网站到本地的完整工作流。当用户说镜像网站、下载网站、抓取网站、本地化网站,或者提供一个网址和保存路径时,必须使用此 skill。输入格式为:镜像网站 www.example.com 保存地址 /a/b/c。自动处理代理、递归下载、文件名清理等全流程。即使用户只是提到想把某个网站保存到本地、离线浏览、克隆站点,... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 128 次。

如何安装 Mirror Website?

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

Mirror Website 是免费的吗?

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

Mirror Website 支持哪些平台?

Mirror Website 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 Mirror Website?

由 wenbing3049(@wenbing3049)开发并维护,当前版本 v1.0.0。

💬 留言讨论