The legal reason
Most countries make unauthorised access to, or interference with, computer systems a criminal offence. The UK has the Computer Misuse Act 1990, the United States has the Computer Fraud and Abuse Act, and EU member states implement Directive 2013/40/EU on attacks against information systems. The details differ, but authorisation is the dividing line in all of them. Testing with the system owner's permission is lawful security work. The same requests without permission may not be, and good intentions are not a defence you want to depend on.
For a scanning service the exposure is direct, because the traffic comes from its infrastructure. If any signed-in user could type in any hostname and start an active scan, the service would be an attack tool with a billing page. Ownership verification is how the service establishes, and can later show, that each active scan was requested by someone with control of the target.
None of this is legal advice. If you test systems that belong to customers or partners, get written authorisation that names the targets and the dates.
The safety reason
Active scanning has side effects even when it is authorised. A scanner submits forms, so it can create records, send emails and open support tickets. It tries credentials, so it can lock accounts. It requests many URLs quickly, so it can trip rate limits, fill logs and wake whoever is on call. Load and resilience tests are designed to stress the target.
The owner of a system can plan for that: choose a window, warn the on-call engineer, point the scan at a test tenant, exclude the "delete account" endpoint. Someone who does not own the system cannot, and the people who do own it get no warning.
Owning the domain is also not always the same as owning the infrastructure. Your marketing site may sit on a hosted platform, and your application behind a content delivery network or on a cloud provider with its own rules about security testing. Proof of ownership tells the scanner you control the name. Checking your providers' testing policies is still your job.
Three common proof methods
All three follow one pattern. The service issues a random token that is tied to the target. You place the token somewhere that only a person in control of the target could. The service then checks for it from the outside.
DNS TXT record. You add a TXT record containing the token to the domain's DNS zone. This proves control of the zone, which is the strongest of the three claims: whoever controls DNS can point the name anywhere. It needs no change to the application and does not affect the running site. It does need access to DNS, which in larger organisations may sit with another team, and new records can take a while to become visible.
File at a well-known path. You serve a small text file containing the token at a fixed URL under /.well-known/ on the target. This proves you can publish content at the root of that host. It suits teams who can deploy but cannot edit DNS. It verifies one host and not the whole domain.
HTML meta tag. You add a meta tag carrying the token to the head of the home page. This proves you can edit the page the host serves at its root. It is the easiest to do on hosted platforms and single-page applications where placing a file at an exact path is awkward.
What makes a verification check trustworthy
The proof is only as good as the check. A careless verifier can be fooled, and the result is a scan of somebody else's system. These are the properties to look for, or to build if you run your own tooling.
1. Fixed locations only. The file must be at the exact well-known path and the meta tag on the root page. If any path will do, a site that hosts user uploads or reflects user input can be "verified" by one of its users.
2. No redirects followed. The token has to be served by the host being verified, and not by wherever that host redirects.
3. Exact token match, with a token that is random, long enough not to be guessed, and specific to the target.
4. Verification bound to the host. If the target's hostname is changed afterwards, the verification no longer applies and has to be done again.
5. Expiry. Domains are sold, projects are handed over and staff leave. A verification from two years ago proves little about today, so it should lapse and be renewed.
6. Private and internal addresses refused, whatever has been verified, so that the scanner cannot be used to reach the service provider's own network.
7. A record of who verified what, by which method, and when.
The gate also has to fail closed. If the service cannot confirm the verification status, for example because a lookup failed, the right answer is to refuse the scan.
What is safe without verification
Not every check is an attack. A passive check makes the same requests an ordinary visitor's browser makes and reads public information: whether the site uses HTTPS, which security headers it sends, what the certificate says. Looking up published vulnerability data for a named software package sends nothing to the target at all. These can reasonably run without proof of ownership, in the same way that anyone can look at a certificate in their browser.
The line is drawn by what is sent. Once a check submits a crafted input to see how the application reacts, such as a quote character in a parameter or a path traversal string, it is active and belongs behind verification, however gentle the probe.
Scans of source code, container images and cloud accounts are a separate case. There the authorisation comes from the connection itself. You granted access to the repository or the account, and nothing is sent to a host that somebody else might own.
How Spoorder handles it
Spoorder supports all three methods: a DNS TXT record, a file at /.well-known/sprdr-verify.txt and a meta tag on the home page. Active engines, including DAST, API security, resilience testing and the AI pentest, will not start without a verified target, and the URL scanned must be on the verified host. Changing a target's host resets its verification, and verification lapses after 90 days. The basic health check and the CVE monitor are passive and run on a registered target without it.
Ready to scan your application?
Free health check: HTTPS, TLS and security headers for one URL, in about a minute. Free account, no credit card.
Scan your app free →