How it works
What ExposureIQ detects, and why it matters
Domains & subdomains
DNS records and certificate-transparency logs, plus a bounded check of common subdomain names (admin, staging, vpn, and similar) that share your domain's wildcard certificate.
Why it matters: Forgotten or undocumented subdomains — an old staging environment, an internal tool someone exposed by accident — are one of the most common ways attackers get a first foothold, precisely because nobody's watching them.
Live hosts & services
Read-only requests to each live host under a domain you have verified: its web ports, a connection to a short fixed list of service ports, and its public site files, confirming what is actually reachable from the public internet. No port-range scanning, no credential guessing. The full list is below.
Why it matters: An asset that resolves in DNS but was never meant to be internet-facing is a different risk than one that's supposed to be public. Knowing which is which is the first question, not an afterthought.
TLS certificates
Certificate subject, issuer, and validity window for every live HTTPS host found.
Why it matters: An expired, mismatched, or unexpectedly-issued certificate is often the first visible sign that something about a host's setup has drifted from what you intended.
Technology & software versions
The specific software and, where identifiable, version running on each asset — a CMS, a web server, a VPN gateway — each with a confidence score for how certain that identification is.
Why it matters: You can't know if you're exposed to a given vulnerability without knowing what you're actually running. This is the fact every other exposure decision downstream depends on.
Whose host it is, and who runs it
Each live host is labelled with how sure the platform is that it belongs to the organization (confirmed, likely, possible or weak) and with who appears to run it: the organization itself, a content network such as Cloudflare, CloudFront or Akamai, a software-as-a-service vendor, or a shared-hosting provider. The labels come from how the host was found, its DNS alias and the network that owns its address, and each one gives its reason in words. They are shown for review. A finding on a host someone else runs is shown with lower confidence, and a host that a software-as-a-service vendor runs is asked only for its home page: its ports, files, administration paths and TLS versions are not probed, because your authorisation of your own domain may not cover another company's platform. The labels do not change the risk score yet.
Why it matters: Treating a vendor's help desk or a CDN's edge as the organization's own server leads to chasing exposures that aren't actually yours to fix. Keeping them separate keeps findings honest, and a person who disagrees with a label can see which rule produced it.
Exposed source maps, secrets & undocumented APIs
On the Advanced tier (and the older Business plans and above), read-only requests for your public-facing JavaScript (same origin only) and API documentation, checked for exposed source maps, hardcoded credentials or API keys, undocumented API endpoints, and sensitive data fields left in public responses.
Why it matters: This is what an attacker reverse-engineers from your website without ever needing privileged access — and it's the category most exposure tools don't look at, because it isn't a CVE.
Known & actively exploited vulnerabilities
Every detected technology correlated against CISA's Known Exploited Vulnerabilities catalog, NVD, and EPSS exploit-prediction scores.
Why it matters: A vulnerability that's already being actively exploited in the wild is a different priority than one that's theoretical. This correlation is what turns a list of software versions into an actual priority order.
Prioritized, evidence-backed findings
Every finding is scored for severity, state (confirmed, version matched, likely, potential, or suspected), and confidence — kept as separate, independent measures — and cites the exact evidence it was created from. "Confirmed" is kept for what a scanner reported or the platform observed directly; a match based on the version a host reports is labelled "version matched", because the vulnerability itself is not tested.
Why it matters: A finding you can't verify isn't actionable, it's noise. Knowing not just what was found but how sure we are, and why, is what lets you trust it enough to act on it.
How discovery actually runs
You confirm you're authorized to scan a domain and prove you control it (a DNS record, a file or a meta tag), then trigger discovery yourself or let scheduled monitoring re-check it. The scheduled run goes through verified domains in turn each night (or less often, if the organization chose a slower schedule), about two minutes for each, so an estate with many domains is covered over several nights. ExposureIQ uses DNS, certificate-transparency and registry data, plus a fixed, bounded set of non-intrusive read-only requests to hosts under domains you have verified: web requests, a connection to six fixed ports, TLS handshakes and public site files. It does not log in, guess credentials, send exploit payloads or scan port ranges. Until a domain is verified, only passive checks run: nothing is sent to your hosts.
What ExposureIQ sends to your systems
This is the complete list, kept in one place in our code and used to write every statement about traffic on this site and in your reports. Requests aimed at your hosts or domain identify themselves with this user-agent:
ExposureIQ-Discovery/0.2 (non-intrusive security assessment; +https://exposureiq.integgreat.in/how-it-works#traffic)
Most hosts see far fewer than the ceiling per run, which is at most 52 requests or connections to one live host on any plan, 70 on Essentials or Starter and 89 on Advanced or Business and above.
| What | Detail | Sent to | Plans | Per host | When |
|---|---|---|---|---|---|
| DNS lookups | DNS lookups for your domain's records (name servers, mail servers, addresses, TXT records, certificate-authority CAA records) and for a bounded list of likely host names (81 common names such as www, mail, vpn, staging, plus up to 25 variations of hosts already found), and two invented names to see whether the domain answers for any name (a wildcard). | DNS resolvers and name servers | All plans | Not per host | Every run, including unverified domains |
| Service and reverse DNS records | DNS lookups for the service records (SRV) your domain may publish, a fixed list of 25 well-known service names such as _sip._tcp and _imaps._tcp, and reverse-DNS lookups for up to 50 of the addresses your domain and its hosts use; both read public DNS, send nothing to your hosts, and only host names under your own domain are kept. | DNS resolvers and name servers | All plans | Not per host | Every run, including unverified domains |
| Dangling DNS alias check | DNS lookups of the alias (CNAME) records of host names that did not resolve and of names that already have a finding, following each alias up to 6 steps and checking whether the name it ends at exists, for up to 150 names per run; it reads public DNS only, sends nothing to your hosts or to the target, and does not try to claim anything. | DNS resolvers and name servers | All plans | Not per host | Every run, including unverified domains |
| Email-authentication DNS records | DNS lookups for SPF, DMARC and eight common DKIM selector names. | DNS resolvers and name servers | All plans | Not per host | Every run, including unverified domains |
| Certificate-transparency search | Searches of public certificate-transparency logs through crt.sh and Cert Spotter (and Chaos where configured) for names issued under your domain, and each night for the certificates issued for your authorised domains (to tell you when a certificate authority new to the domain issues one); your domain name is sent to those services, nothing is sent to your hosts. | Third-party services | All plans | Not per host | Every run, including unverified domains |
| Registry (RDAP) data | Registration (RDAP) lookups for your domain and for the network that owns each address found. | Third-party services | All plans | Not per host | Every run, including unverified domains |
| Zone-transfer request | One standard DNS zone-transfer request (AXFR) to each of up to four of your name servers on port 53, which succeeds only where a server is misconfigured to hand its zone to anyone. | Your hosts | All plans | Not per host | Verified domains only |
| Web request to each host | One read-only GET of the home page of each host that exists, on port 443 and, if that does not answer, on port 80 (certificate problems are recorded, not refused); for a host that has both an IPv4 and an IPv6 address, the same request to one address of each family, to compare how they answer; for a domain that answers for any name, also one GET of an invented name, once per run, to recognise the wildcard page. | Your hosts | All plans | Up to 4 | Verified domains only |
| Fixed service ports | One connection each to ports 22, 25, 53, 1433, 3306, 3389, 5432, 6379, 9200, 11211 and 27017 (remote access, mail, DNS and common database, cache and search services) to see whether they are open (no data is sent and no login is attempted; on ports 22 and 25 the one line of greeting that an SSH or mail server sends by itself is also read, at most 256 bytes, to learn which software it is), and one read-only GET of the home page on each of ports 8080 and 8443. | Your hosts | All plans | Up to 13 | Verified domains only |
| Files that should not be public | One read-only GET of each of 13 fixed paths where a backup, a database dump, a settings file with passwords or a source-code folder is sometimes left public (such as /.env, /.git/HEAD and /backup.sql), on the port the host answered on, four at a time; at most the first few kilobytes of each answer are read, only to test them against what that kind of file looks like, and none of it is stored. | Your hosts | All plans | Up to 13 | Verified domains only |
| Administration and monitoring interfaces | One read-only GET of each of 6 fixed paths where the page of a database manager, a build server or a monitoring server is sometimes left public (such as /phpmyadmin/ and /login), on the port the host answered on, four at a time; at most the first few kilobytes of each answer are read, only to test them against what that product's page looks like, none of it is stored, and no login is ever attempted. | Your hosts | All plans | Up to 6 | Verified domains only |
| Public site files | Read-only GETs of /.well-known/security.txt (and /security.txt if absent), /robots.txt and /sitemap.xml. | Your hosts | All plans | Up to 4 | Verified domains only |
| TLS and HTTP baseline | Up to five TLS handshakes with each HTTPS host (one normal and one for each of TLS 1.0, 1.1, 1.2 and 1.3, to learn which protocol versions are accepted), up to six TLS connection attempts offering only weak cipher families (RC4, 3DES, DES, export, NULL, anonymous; these complete no session), and one GET of the plain-HTTP page on port 80 without following redirects. | Your hosts | All plans | Up to 12 | Verified domains only |
| Product identification | Up to 18 fixed read-only GETs of version or banner pages of specific infrastructure and AI-serving products on their usual ports (11434, 4000, 8000, 3000, 8080, 2375, 6443, 5000, 9000, 9443, and two paths on the web ports); it never lists models, sends prompts, calls a chat or completion route, or uses a credential. | Your hosts | Essentials, or Starter and above | Up to 18 | Verified domains only |
| Public JavaScript and API documentation | Per web host, up to 5 public JavaScript files referenced by the home page, up to 5 of their source maps, 6 fixed API-documentation paths (/swagger.json, /openapi.json and similar) and up to 3 API endpoints named in your own JavaScript, each fetched with a read-only GET. | Your hosts | Advanced, or Business and above | Up to 19 | Verified domains only |
| Cloud storage name checks | Up to 30 read-only requests to Amazon S3, Google Cloud Storage and Azure Blob Storage for bucket names derived from your domain and organization name, to see whether any is publicly listable; these go to the cloud providers, not to your hosts. | Third-party services | Advanced, or Business and above | Not per host | Verified domains only |
| Domain-control check | When you press Check now to prove you control a domain, one DNS TXT lookup, or one GET of /.well-known/exposureiq-challenge.txt or of the home page, depending on the method you chose; the check is repeated about monthly for verified domains. | Your hosts | All plans | Up to 1 | Only when you ask |
| Screenshots | When you ask for a screenshot of an asset, its URL is sent to the screenshot service ScreenshotOne, which loads the page in a real browser from its own network (including the page's own images, scripts and styles) and returns the image. | Third-party services | Essentials, or Starter and above | Not per host | Only when you ask |
| Fix verification | When you press Verify on a remediation, one DNS lookup for that one host and then only what that finding needs: for a vulnerability finding the same home-page GET as the web request above; for a configuration finding only the checks its rule reads: a protocol-version rule up to five TLS handshakes, the weak-cipher rule one handshake and six weak-cipher attempts, the redirect rule one handshake and one plain-HTTP GET, a certificate rule one handshake, a header or cookie rule one or two web requests; never all of them. | Your hosts | All plans | Up to 7 | Only when you ask |
Plan names: All plans; Essentials, or Starter and above; Advanced, or Business and above. Nothing here logs in, guesses a password, sends an exploit payload or scans a range of ports.