MCP documentation menu

Domain reconnaissance

Gathering intelligence on one domain, then pivoting outward from known domains to find related ones. Each phase names the tool and what to read in its answer; the response formats are in the tools’ own topics.

Investigating one domain

Phase 0: Overview and health

# Scrape + DNS + RDAP in one call, with a summary of the key facts
datapulse_domain_overview(domain="example.com")

# Live posture probe: delegation, DNSSEC, nameservers, TLS, HTTP/HSTS, mail, CAA/TLSA
datapulse_live_domain_health(domain="example.com")

The overview answers most first questions (does it resolve, who is the registrar, what is the site); the health probe gives a verdict (ok, errors, warnings) on how the domain is run. See datapulse_help(topic="scraping") and datapulse_help(topic="domain_health").

Phase 1: DNS

datapulse_live_dns(domain="example.com")

The default lookup covers the main record types and the _dmarc TXT record (the list and what the summary strips: “Default lookup” in datapulse_help(topic="dns")). The summary shows:

  • NS → the DNS provider
  • MX, TXT (SPF) and dmarc → mail posture (datapulse_help(topic="email_security"))
  • CNAME chains → CDN use
  • CAA → which CAs may issue certificates
  • dnssec_signed → whether answers validated

DNSSEC chain-of-trust inspection and the full response format are in datapulse_help(topic="dns").

Phase 2: Registration

datapulse_live_rdap(query="example.com")

It gives the creation and expiry dates, the registrar (ianaid), the EPP status array (locks and holds) and the nameservers. live_rdap can answer from a cached summary; the envelope’s source says which (cache, datapulse, live), and force=true asks the registry afresh.

Age measures how long the string has existed, not who registered it or holds it now; the history in Phase 3 sets it in context. A registrar name says little about the owner (acquisitions, resellers, regional offices); lock depth compared with a domain the owner certainly holds says more (datapulse_help(topic="lookalike_domains")). Status codes and registrant data: datapulse_help(topic="rdap").

Phase 3: History

datapulse_dns_dphistory(domain="example.com")        # nameserver, MX and web-hosting providers over time
datapulse_dns_rdap_history(domain="example.com")     # registrar transfers and abuse-contact changes

The hosting history records nameserver and MX hostnames with ASNs (numbers only, no operator names) and the web-hosting ASN, as a change log at mostly month granularity since 2023-04-07. It shows provider migrations, ASN changes and recent changes. Its a/c/d entries are DataPulse listing events, not registry events. See datapulse_help(topic="history").

Phase 4: Infrastructure similarity

datapulse_dns_dptechsim(domain="example.com")

Domains with a similar nameserver, mail and hosting stack, most similar first, at the default threshold 0.95. A high score means a similar stack, not a common operator: the matches are candidates to examine. Score bands and the result limit: datapulse_help(topic="infrastructure").

Identifying providers

DNS (from NS)

PatternProvider
*.ns.cloudflare.comCloudflare
*.awsdns-*.com/net/orgAWS Route 53
ns-cloud-*.googledomains.comGoogle Cloud DNS
ns1.google.com … ns4.google.comGoogle’s own zones (google.com, golang.org), not a customer DNS service
*.azure-dns.comAzure DNS
*.domaincontrol.comGoDaddy

Mail (from MX)

PatternPlatform
aspmx.l.google.com, altN.aspmx.l.google.com, smtp.google.comGoogle Workspace (Gmail)
*.mail.protection.outlook.comMicrosoft 365
*.pphosted.comProofpoint
*.mimecast.comMimecast

Matching the MX suffix identifies the platform, not the account holder: an MX can name a platform that never provisioned the domain (datapulse_help(topic="email_security")).

Hosting and CDN (from CNAME and addresses)

PatternProvider
CNAME *.cloudfront.netAWS CloudFront
CNAME *.fastly.netFastly
CNAME *.akamaiedge.netAkamai

A bare address names its holder only through the RIR record. First octets are split among many organisations, so mapping a prefix to a provider by eye misleads. datapulse_live_rdap(query="<ip>") returns the RIR’s network_name and, where the RIR publishes it, organization. The hosting history gives the ASN number of the web address; the RIR record names the network.

Subdomains and services

Common hostnames, each a separate lookup:

datapulse_live_dns(domain="www.example.com")
datapulse_live_dns(domain="mail.example.com")
datapulse_live_dns(domain="vpn.example.com")
datapulse_live_dns(domain="remote.example.com")

A random label answering (datapulse_live_dns(domain="randomstring12345.example.com")) means a wildcard; the health probe tests this with three random labels (wildcard.status).

Services by SRV record:

datapulse_live_dns(domain="_sipfederationtls._tcp.example.com", record_type="SRV")   # Microsoft Teams federation
datapulse_live_dns(domain="_xmpp-server._tcp.example.com", record_type="SRV")        # XMPP
datapulse_live_dns(domain="_ldap._tcp.example.com", record_type="SRV")               # Active Directory

Pivoting outward from known domains

Starting from domains an organisation certainly owns, the same tools surface possible related domains: subsidiaries, campaign microsites, vendor-run portals, or domains set up outside normal IT control (shadow IT). They show related infrastructure and registration patterns; they do not show ownership or whether a domain is authorised. The output is a shortlist for the organisation to confirm.

# 1. Profile a known domain
datapulse_domain_overview(domain="example.com")

# 2. Infrastructure siblings of the known domain (default threshold 0.95)
datapulse_dns_dptechsim(domain="example.com")

# 3. Profile each promising match: branding, entity, mail and hosting, registration dates
datapulse_domain_overview(domain="candidate-example.com")

# 4. Resolve registrar_iana_id where it is an IANA ID (it is the registry-published number:
#    infrastructure topic, "Registrar identifiers")
datapulse_dns_registrar(iana_id=146)

datapulse_live_dns, datapulse_live_rdap and datapulse_dns_dphistory on a candidate give the records, the registration and the timeline when the overview is not enough.

Signals worth a closer look:

  • the candidate’s site names the organisation, a product, a department or an event
  • mail or hosting matching the organisation’s normal providers
  • a recent registration on infrastructure close to established assets
  • a registrar different from the organisation’s usual portfolio pattern, such as a consumer-oriented retail registrar where the portfolio sits at an enterprise one; this is a triage cue, since subsidiaries and acquired brands keep their own registrars

Noise and limits:

  • Shared platforms (Cloudflare, AWS, Google Cloud, Shopify) create many unrelated matches.
  • A registrar on its own is weak evidence.
  • Similarity suggests related infrastructure, not confirmed ownership.
  • Scraped content can reflect a vendor, a reseller or a parking page rather than the organisation.

Typical requests of this kind: “find possible shadow IT for example.com”, “look for unmanaged domains related to our primary domains”.

Generated from the live server (DataPulse MCP v0.9-341-g27163c2) on October 20, 2026. Your AI assistant reads this page by calling datapulse_help(topic="reconnaissance").