Mcp Outbound Infrastructure Fingerprint

com.mambabuilt由 mambabuilt.com 發布v1.3.1更新於 Nov 7, 2026

Detect whether a company runs cold outbound, on what stack, and from which sending domains.

已驗證STDIO僅桌面網頁搜尋與抓取商業與電商

概覽

AI 產生的概覽

本機 MCP 伺服器，透過 Apify actor 依公司網域判斷其是否在做冷郵件開發，以及使用什麼技術堆疊。

可以做什麼
  • 針對單一公司網域或一批網域進行冷郵件開發特徵分析
  • 回傳 runs_outbound 判定（program、light、none、unknown），並附信心度與可引用的證據
  • 報告寄送網域、註冊叢集、寄送平台、收件服務商、暖機情況與基礎設施供應商
  • 檢查 SPF、DKIM 與 DMARC 設定，可選另計費的可達性與黑名單檢查
  • 可調整掃描深度、探測數量上限、逾時、略過快取以及單次呼叫的費用上限
適用情境
適合業務、GTM 或研究情境：需要判斷潛在客戶或競爭對手是否真的在做冷郵件開發、使用哪套寄送堆疊和哪些網域，而不必手動排查 DNS。
執行需求
透過 npx 以 stdio 在本機執行，需要 Node.js。需要 Apify 帳號與 APIFY_TOKEN 環境變數，底層 actor 的費用計入該 Apify 帳號。呼叫 actor 需要網路連線。
安裝前請注意
每次執行都會消耗 Apify 額度：包含啟動費與依網域計費，即使沒有開發訊號或網域無法解析也照樣計費。可選的可達性檢查另行計費，可用 max_charge_usd 限制花費。APIFY_TOKEN 屬於機密，不應外流。結果為啟發式判斷：品牌詞為常見英文單字時信心度會被壓低，寄送平台的辨識率也只是部分涵蓋。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Mcp Outbound Infrastructure Fingerprint，將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Desktop only，透過 STDIO。 STDIO 服務會啟動本機處理程序，因此需要 SourceWeft 桌面主機。

其他 MCP 客戶端

參照 儲存庫 中的啟動說明。

README

Outbound Infrastructure Fingerprint MCP Server

[npm] [MCP] [License: MIT]

[Outbound Infrastructure Fingerprint MCP server]

MCP server for the Mamba Labs Outbound Stack Detector actor on Apify.

Give it a company domain. It tells your agent whether that company runs cold email outbound, and on what stack.

Tool

fingerprint_outbound_infrastructure

InputTypeNotes
domainstringOne company domain.
domainsstring[]Batch. Takes precedence over domain.
scan_sending_domainsbooleanDefault true. The strongest signal and the slowest step.
sending_domain_depthdeep | standardDefault deep (.com .co .io .net .org).
check_deliverabilitybooleanDefault false. Adds a separately billed blacklist check and health score.
skipCachebooleanIgnore the 7 day result cache.
max_sending_domain_probesinteger10 to 300, default 80. Cap on candidate sending domains probed per company.
request_timeout_msinteger3000 to 20000, default 9000. Per HTTP request timeout.
dns_timeout_msinteger1000 to 15000, default 4000. Per DNS lookup timeout.
max_charge_usdnumberOptional spending cap for this call, in USD. The actor stops before any charge that would take the run past this amount and keeps every row already charged.

Returns a flat row per domain: runs_outbound (program, light, none, unknown), confidence, sending_domains[], registration_clusters[], sending_platforms[], inbox_provider, warmup_detected, infrastructure_vendors[], spf_status, dkim_status, dmarc_policy, and an evidence[] array of quotable strings.

What it actually detects

A real outbound program does not send from the domain the business runs on. It buys lookalikes (getcompany.com, company-mail.com, trycompany.co), gives them their own mail tenant, and redirects their web root at the real site. That redirect is the attribution handle, and finding those domains is the part nobody else does.

Measured on nine live domains: 6 of 6 detected on companies that demonstrably run outbound, 0 false positives of 3 on controls.

Two honest limits. When the brand token is an ordinary English word (Gong, Clay, Ramp), lookalike patterns collide with unrelated businesses that legitimately own them; the actor sets brand_is_common_word and caps confidence rather than reporting a confident none. And sending-platform recall is partial by design: Instantly, Smartlead, Lemlist, Apollo and Salesloft connect over OAuth to a customer's own mailbox and publish no SPF include host at all, so an empty sending_platforms tells you little while a populated one is solid.

Pricing

You pay per event on Apify, billed to your own Apify account. Prices from the live pricing record, read 2026-10-16:

EventPrice per event (USD)Fires when
actor-start$0.00005Charged once when the run starts.
domain-analyzed$0.004 on the Free plan, down to $0.0034 on higher Apify plansCharged once per domain analyzed. Domains that return no outbound signals, or that do not resolve, still count as analyzed and are charged.

The actor's Apify page carries the current prices. Running this server consumes Apify credits.

Setup

json
{  "mcpServers": {    "mamba-outbound-infrastructure-fingerprint": {      "command": "npx",      "args": ["-y", "@mambalabsdev/mcp-outbound-infrastructure-fingerprint"],      "env": { "APIFY_TOKEN": "your-apify-token" }    }  }}

Get a token at console.apify.com/account/integrations. Read-only; consumes Apify credits per domain analyzed.

Also available

This tool is also exposed by the GTM Suite umbrella server, alongside the rest of the Mamba Labs GTM actors, if you would rather run one server than many.

Built by Mamba Labs.

來源：README.md，提交 55e1163

工具

1

fingerprint_outbound_infrastructure

Fingerprint Outbound Infrastructure

Given a company domain, determine whether that company runs cold email outbound and on what stack. Returns a runs_outbound verdict of program (a deliberate cold outbound setup), light (one weak signal), none, or unknown, plus the evidence behind it. The strongest signal is the lookalike sending domains a real outbound program leaves behind: domains like getcompany.com or company-mail.com that carry their own mail and redirect back to the primary site. Also returns the inbox provider (Google Workspace, Microsoft 365 and others) for the primary domain and each sending domain, any detected sending platform (Outreach, Salesloft, Lemlist, Instantly, Smartlead, Apollo and more), registration clusters showing sending domains bought on the same day, cold email infrastructure vendors, and deliverability posture (SPF, DKIM, DMARC). Note that sending platform recall is partial by design: sequencers that connect over OAuth to a customer's own mailbox leave no DNS trace, so an empty sending_platforms means little while a populated one is solid. Public DNS and HTTP redirects only, no login, no mailbox access. Returns flat Clay-ready JSON. Reads public data and writes nothing; requires an APIFY_TOKEN and consumes Apify credits per domain analyzed.

輸入參數domaindomainsskipCachedns_timeout_msmax_charge_usdrequest_timeout_mscheck_deliverabilityscan_sending_domainssending_domain_depthmax_sending_domain_probes

版本歷史

5
  1. v1.3.1最新Nov 5, 2026
  2. v1.3.0Oct 16, 2026
  3. v1.2.1Oct 6, 2026
  4. v1.2.0Oct 6, 2026
  5. v1.1.1Sep 16, 2026

評分

已滿足適用檢查項的 90%

該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通25 分
  • 已宣告認證方式10 分

    請在註冊中心條目中宣告憑證請求標頭，或保持端點能夠回應探測，以便判定其認證方式。

  • 已列出工具必需15 分
  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分