← 返回 Skills 市场
2.1 修改
2.2 修改
Auto Building
作者
addogiavara-tech
· GitHub ↗
· v1.0.0
485
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install auto-building
功能描述
基于 AUTO-BUILDING 源码,生成定制化信息采集管理系统。用户指定栏目、分类、数据源后,系统自动采集、审核、发布内容。
使用说明 (SKILL.md)
AUTO-BUILDING
基于 AI 的自动化信息采集管理系统技能。
什么是 AUTO-BUILDING
AUTO-BUILDING 是一个开源的信息采集管理平台,可以帮助用户快速构建定制化的内容采集系统。用户只需指定想要采集的栏目、分类和数据源,系统会自动完成采集、审核、展示的全流程。
何时使用此技能
当用户想要:
- 构建自己的新闻早报系统
- 监控系统监控竞品价格和动态
- 采集特定行业资讯(如科技、财经、医疗)
- 建立资源导航页面(如 AI 工具集、开源项目)
- 定制化信息聚合平台
快速开始
步骤 1:克隆源码
git clone https://github.com/hasd52636-a11y/Auto_Building_new
cd Auto_Building_new
步骤 2:安装依赖
npm install
步骤 3:启动服务
npm run dev
步骤 4:访问
- 首页:
http://localhost:3000 - 管理后台:
http://localhost:3000/admin
如何根据用户需求生成系统
步骤 1:理解用户需求
当用户提出需求时,识别:
- 目标栏目(如:新闻早报、产品监控、行业资讯)
- 分类规则(如:科技/财经/体育、亚马逊/淘宝/京东)
- 数据来源(如:36kr、虎嗅、GitHub 仓库、RSS 订阅)
步骤 2:修改配置文件
根据用户需求修改以下文件:
2.1 修改 config/sources.json
{
"primaryCategories": ["智能母体", "用户想要的栏目"],
"secondaryCategories": {
"智能母体": ["agent", "skill", "mcp", "prompt", "opensource_soft"],
"用户想要的栏目": ["分类1", "分类2", "分类3"]
},
"sources": [
{
"name": "数据源名称",
"type": "github|rss|directory|custom",
"url": "地址",
"enabled": true
}
]
}
2.2 修改 src/data/resources.ts
export const PRIMARY_CATEGORIES = [
'智能母体',
'用户想要的栏目',
];
export const SECONDARY_CATEGORIES = [
{ labelKey: 'type.agent', value: 'agent' },
{ labelKey: 'type.skill', value: 'skill' },
{ labelKey: 'type.分类1', value: 'category1' },
{ labelKey: 'type.分类2', value: 'category2' },
];
步骤 3:指导用户配置
告诉用户执行以下操作:
- 运行
npm install && npm run dev - 访问
http://localhost:3000/admin - 在「采集规则」Tab 添加具体采集规则
- 在「数据源」Tab 确认数据源已添加
- 点击「执行采集」开始采集内容
- 在「审核」Tab 审核采集的内容
- 内容自动展示在首页
配置说明
数据源类型
| 类型 | 说明 | 示例 |
|---|---|---|
github |
GitHub 仓库 | composiohq/awesome-claude-skills |
rss |
RSS 订阅 | https://36kr.com/feed/ |
directory |
目录网站 | https://skills.sh/ |
custom |
自定义 URL | 任意网页地址 |
采集规则
用户可通过管理后台 /admin 的「采集规则」Tab:
- 手动添加正则匹配规则
- 使用 AI 自然语言生成规则
- 设置关键词过滤
- 设置内容类型(视频/图文/全部)
示例对话
示例 1:新闻早报系统
用户:"帮我做个科技新闻早报,每天采集36kr和虎嗅"
修改内容:
config/sources.json:
{
"primaryCategories": ["智能母体", "新闻早报"],
"secondaryCategories": {
"新闻早报": ["科技", "财经", "创投"]
},
"sources": [
{
"name": "36kr",
"type": "rss",
"url": "https://36kr.com/feed/",
"enabled": true
},
{
"name": "虎嗅",
"type": "rss",
"url": "https://www.huxiu.com/rss",
"enabled": true
}
]
}
告诉用户:
系统已配置完成!运行以下命令启动:
cd Auto_Building_new
npm install
npm run dev
然后访问 http://localhost:3000/admin
1. 在「采集规则」添加:只采集今天发布的文章
2. 点击「执行采集」
3. 在「审核」Tab 审核内容
示例 2:产品监控系统
用户:"做个电商产品价格监控系统,监控亚马逊和淘宝"
修改内容:
config/sources.json:
{
"primaryCategories": ["智能母体", "产品监控"],
"secondaryCategories": {
"产品监控": ["亚马逊", "淘宝", "京东"]
},
"sources": [
{
"name": "亚马逊",
"type": "directory",
"url": "https://www.amazon.cn",
"enabled": true
},
{
"name": "淘宝",
"type": "directory",
"url": "https://www.taobao.com",
"enabled": true
}
]
}
示例 3:AI 工具集
用户:"我想做一个 AI 编程工具导航页面"
修改内容:
config/sources.json:
{
"primaryCategories": ["AI编程工具"],
"secondaryCategories": {
"AI编程工具": ["代码补全", "代码审查", "自动化测试"]
},
"sources": [
{
"name": "GitHub Copilot",
"type": "github",
"repo": "github/copilot",
"enabled": true
},
{
"name": "Cursor",
"type": "github",
"repo": "getcursor/cursor",
"enabled": true
}
]
}
核心功能(不可修改)
- 采集脚本:
scripts/scrape-all.ts,scripts/classify.ts - 审核流程:
pending-review.json→approved-resources.json - 展示极面:动态读取配置渲染栏目和分类
注意事项
- 只修改配置文件:
config/sources.json和src/data/resources.ts - 不要修改采集脚本和页面逻辑
- 运行后需要执行
npm run daily采集内容 - 在
/admin审核采集的内容后才会展示在首页 - 如果需要更多自定义,可以通过管理后台的「采集规则」和「数据源」Tab 进行配置
安全使用建议
This skill description and instructions are coherent for building a scraping/aggregation site, but the package itself is instruction-only and points you to clone and run a GitHub repository. Before installing or running it: 1) Inspect the repository (https://github.com/hasd52636-a11y/Auto_Building_new) — review package.json, scripts, and any server-side code for unexpected network calls, credential handling, or malware. 2) Run npm install and the app in a sandbox or isolated environment (container/VM). 3) Review dependencies for known vulnerabilities and check whether the app asks for credentials or modifies system files. 4) Consider legal/terms-of-service issues for scraping target sites and respect robots.txt. If you want a higher-assurance verdict, provide the upstream repository contents so they can be statically inspected; currently that runtime code was not included in this package.
功能分析
Type: OpenClaw Skill
Name: auto-building
Version: 1.0.0
The skill is classified as suspicious due to a significant supply chain vulnerability. The `SKILL.md` file instructs the OpenClaw agent to clone an external GitHub repository (`https://github.com/hasd52636-a11y/Auto_Building_new`) and then execute `npm install` and `npm run dev` on the cloned code. While the instructions themselves do not exhibit malicious intent, this process allows for the execution of arbitrary code from an external, unverified source, posing a remote code execution risk if the upstream repository or its dependencies are compromised. This falls under 'risky capabilities without clear malicious intent' as per the classification guidelines.
能力评估
Purpose & Capability
Name/description (auto content-aggregation) match the declared requirements: Node/npm are required and the SKILL.md instructs cloning and running a Node app. Nothing in the skill metadata requests unrelated credentials, binaries, or config paths.
Instruction Scope
SKILL.md explicitly directs the agent/user to clone https://github.com/hasd52636-a11y/Auto_Building_new, run npm install, and run the app; it also instructs editing config/sources.json and running scraping tasks. These actions are consistent with the stated purpose but involve executing code fetched from a remote repository and initiating web scraping, so the remote repository's contents determine actual runtime behavior and risk.
Install Mechanism
There is no bundled install payload in the skill package; the instructions point to a GitHub repo (well-known host) to clone. Using GitHub is expected for this purpose, but because the skill does not include the repo code, the reviewer could not inspect the actual runtime code. The SKILL.md also contains an 'install' clone entry referencing the same GitHub repo — consistent but worth verifying.
Credentials
No environment variables or credentials are requested by the skill package. The described functionality (scraping external sources) normally does not require platform secrets. Beware that the cloned repository could itself request or use credentials at runtime (not visible in this package).
Persistence & Privilege
Skill is not marked always:true and uses normal, user-invocable/autonomous defaults. It does not request elevated platform persistence or modify other skills. Running the upstream app may require setting up scheduled jobs (e.g., npm run daily), but that's standard for this kind of application.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install auto-building - 安装完成后,直接呼叫该 Skill 的名称或使用
/auto-building触发 - 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: Automated content aggregation system based on AUTO-BUILDING source code
元数据
常见问题
Auto Building 是什么?
基于 AUTO-BUILDING 源码,生成定制化信息采集管理系统。用户指定栏目、分类、数据源后,系统自动采集、审核、发布内容。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 485 次。
如何安装 Auto Building?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install auto-building」即可一键安装,无需额外配置。
Auto Building 是免费的吗?
是的,Auto Building 完全免费(开源免费),可自由下载、安装和使用。
Auto Building 支持哪些平台?
Auto Building 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 Auto Building?
由 addogiavara-tech(@addogiavara-tech)开发并维护,当前版本 v1.0.0。
推荐 Skills