Apify Site audit

io.github.AHMEDFOUAAD由 GitHub 账号 AHMEDFOUAAD 发布v1.0.0更新于 Nov 11, 2026

Sitemap URL extraction, bulk URL status / broken-link checks and dataset dedup via Apify.

发布方自己的服务:audit.domains - Domain Valuation

已验证Streamable HTTP可网页运行网页搜索与抓取数据与分析

概览

AI 生成的概览

由 Apify 托管的 MCP 预设,为助手提供三个公开网页站点审计工具:站点地图 URL 提取、批量链接状态检查与数据集去重。

可以做什么
  • 从站点地图中提取 URL 并输出结构化 JSON
  • 批量检查大量 URL 的状态码与失效链接
  • 对结果数据集进行去重与合并
  • 通过 Apify 标准的运行与数据集辅助工具查看和管理生成的 Actor 运行与数据集
适用场景
适合让助手审计站点地图与链接,或清理抓取到的 URL 列表,而无需加载整个 Apify Store 工具集。适用于可接受按结果计费的周期性公开网页检查。
运行要求
使用 mcp.apify.com 远程端点并带上该预设的 tools 参数。需要 Apify 账号与身份验证:OAuth 登录,或在 Authorization 请求头中以 Bearer 形式提供 Apify API 令牌。运行费用计入你自己的 Apify 账号。
安装前请注意
运行按结果计费并计入你的 Apify 账号(每次运行有一定免费结果额度,免费套餐含有限月度额度)。请妥善保管 Apify 令牌,切勿提交到代码库。这些 Actor 由仓库作者发布并销售,与 Apify 无隶属关系。仅读取公开页面。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Apify Site audit,将其添加到工作区。
  2. 连接 所需的 凭据。它们会被私密存储,绝不会在公开页面上展示。
  3. 为需要使用其工具的对话启用该服务。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体,或 SourceWeft 对话中。智能体会读取该服务器的安装指南,先向你展示发布者、端点和登录方式,经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.AHMEDFOUAAD%2Fapify-site-audit/install.md,并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前,先告诉我它的发布者、端点和登录方式,等我确认。不要把 API 密钥写进命令或我会提交的文件;问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。 客户端会在首次连接时完成 OAuth 登录。

{
  "mcpServers": {
    "apify-site-audit": {
      "type": "http",
      "url": "https://mcp.apify.com/?tools=ahmed.fouaad82/sitemap-extractor,ahmed.fouaad82/url-status-checker,ahmed.fouaad82/dataset-dedup-merge"
    }
  }
}

README

Apify MCP data presets for AI agents

Copy-paste MCP configs that give Claude Desktop, Cursor, VS Code (Copilot) or Claude Code a small, focused set of public-web-data tools instead of the whole Apify Store. Each bundle points the official hosted Apify MCP server at three specific Actors using its documented ?tools= URL parameter, so the agent sees exactly three dedicated tools plus the run/dataset helpers.

Disclosure: I (GitHub @AHMEDFOUAAD, Apify user ahmed.fouaad82) built and sell these Actors on Apify Store. They are paid per result (pay-per-event). The first results of every run are free (25 for most of these Actors; see each Store page's Pricing tab), then the listed price applies. You need your own Apify account (OAuth sign-in or your own API token); runs bill to your account, and Apify's free plan includes $5/month of usage. This repo only contains client configs: it is not affiliated with or endorsed by Apify.

BundleActorsEndpoint
Public job boards (Greenhouse, Lever, multi-ATS)greenhouse-public-jobs, lever-salary-jobs, ats-public-jobshttps://mcp.apify.com/?tools=ahmed.fouaad82/greenhouse-public-jobs,ahmed.fouaad82/lever-salary-jobs,ahmed.fouaad82/ats-public-jobs
Content for RAG (RSS, podcast transcripts, PDF text)rss-feed-reader, podcast-published-transcripts, public-pdf-texthttps://mcp.apify.com/?tools=ahmed.fouaad82/rss-feed-reader,ahmed.fouaad82/podcast-published-transcripts,ahmed.fouaad82/public-pdf-text
Site audit (sitemap URLs, link/status checks, dedup)sitemap-extractor, url-status-checker, dataset-dedup-mergehttps://mcp.apify.com/?tools=ahmed.fouaad82/sitemap-extractor,ahmed.fouaad82/url-status-checker,ahmed.fouaad82/dataset-dedup-merge

Each bundle folder has: cursor.mcp.json, vscode.mcp.json, token variants, claude_desktop_config.json (stdio bridge via mcp-remote) and a server.json (official MCP Registry entry).

Why presets?

  • Fewer, sharper tools. A preselected endpoint exposes only the tools for one job, which keeps the agent's tool list short and its choices predictable (Apify recommends pinning tools in production).
  • Fresh public data. Job boards, feeds, transcripts, PDFs, sitemaps and link status, fetched at run time, returned as structured JSON.
  • Pay per result. No subscription for these Actors; the first results of every run are free.

Verified

On 2026-11-11 the three endpoints were initialized against https://mcp.apify.com (server apify-mcp-server 0.21.5) and tools/list returned the expected tools, for example for the jobs bundle:

text
ahmed-dot-fouaad82--greenhouse-public-jobsahmed-dot-fouaad82--lever-salary-jobsahmed-dot-fouaad82--ats-public-jobsget-actor-run, get-dataset-items, get-key-value-store-record, abort-actor-run

The jobs bundle's claude_desktop_config.json was also run as-is through npx mcp-remote@latest (stdio bridge, Node 22): initialize and tools/list returned the same 7 tools.

A real tool call (ahmed-dot-fouaad82--rss-feed-reader, 3 items from the Hacker News RSS feed) returned SUCCEEDED with 3 dataset rows and 0 charged results (free slice). Re-run the check yourself with scripts/check_preset.py.

Notes from that test:

  • Endpoints that preload an Actor require authentication (OAuth or token); without it the server answers HTTP 401. Only discovery tools (search-actors, fetch-actor-details, docs) work anonymously.
  • OAuth sign-in was not exercised by the script (it used a token header); OAuth is the method Apify documents as recommended.

Security

  • Never commit your Apify token. The token variants read it from an environment variable or a secure prompt.
  • The Actors only read public pages (no logins, no private data). Each Store page documents its sources and robots/ToS limits.

Licence

MIT for the configs and script in this repo. The Actors are separate paid products on Apify Store.

来源:README.md,提交 e6e1639

工具

0
该服务器需要登录,因此 SourceWeft 无法在没有账号的情况下读取其工具,也未收录。已连接账号所能看到的内容可能因账号及其权限而异。

版本历史

1
  1. v1.0.0最新Nov 11, 2026

评分

已满足适用检查项的 85%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分

其他 Audit 集成