Skip to content
AdminKit

How do you read an HTTP security headers check result?

AdminKit's HTTP security headers result lists each recommended response header — CSP, Strict-Transport-Security, X-Content-Type-Options, X-Frame-Options or frame-ancestors, Referrer-Policy and Permissions-Policy — as present or missing, with the exact value sent for each one present, because presence alone doesn't say whether the policy it carries is strict enough to do its job.

By AdminKit team, Netbase JSC · Updated

Key takeaways

  • The result is a checklist for one fetch of one URL. Each of six headers is marked present or missing, and the exact value is shown for each one that is present.
  • Findings carry a severity. A missing HSTS header (on an https:// URL) or a missing clickjacking defence is medium, a missing CSP or nosniff is low, and a missing Referrer-Policy or Permissions-Policy is info.
  • “Present” means sent, not strong. A Content-Security-Policy of default-src * is present and does almost nothing.
  • Fix in this order: Content-Security-Policy and Strict-Transport-Security first, then nosniff, clickjacking protection and Referrer-Policy.
  • Headers are one browser-side defence layer. A clean result does not mean the application has no other vulnerabilities.

What is the result actually reporting?

AdminKit’s HTTP security headers check fetches a URL and reads back its response headers, the same way a browser’s first request to a page does. The result is a checklist, not a single score: each recommended header is marked present or missing, and every present header shows the exact value the server sent. These security headers all work the same general way — they’re instructions a server gives a browser about how to treat the page it’s about to render — and a browser only enforces an instruction it actually received. A header that’s missing isn’t silently “defaulted to safe”; the related protection simply doesn’t apply at all.

The check follows redirects, so the headers you see belong to the final response. When the final URL differs from the one you entered, the result shows it. That matters: an http:// URL that redirects to https:// is judged on the HTTPS response’s headers.

Part of the resultWhat it tells you
Final URL after redirectsWhich response the headers came from
Six header rowsContent-Security-Policy, Strict-Transport-Security, X-Content-Type-Options, X-Frame-Options, Referrer-Policy, Permissions-Policy: present with value, or missing
Findings listEach gap, with a severity and a one-line fix hint

Which findings can the check raise, and how serious are they?

Each finding has a fixed trigger. Knowing the trigger explains results that look odd at first, such as an HSTS finding that appears for one URL and not another.

FindingSeverityRaised whenFix hint shown
No Strict-Transport-Security headermediumThe URL is https:// and the header is absentmax-age=31536000; includeSubDomains
No clickjacking protectionmediumNo X-Frame-Options and no CSP frame-ancestorsX-Frame-Options: DENY or frame-ancestors
No Content-Security-Policy headerlowHeader absentRestrict script, style and frame sources
No X-Content-Type-Options: nosnifflowHeader absent, or its value is not nosniffX-Content-Type-Options: nosniff
No Referrer-Policy headerinfoHeader absentstrict-origin-when-cross-origin or stricter
No Permissions-Policy headerinfoHeader absentRestrict unused browser features
Server header discloses versioninfoServer contains a version, such as nginx/1.2Remove or generalise the version

Two details to notice. The HSTS finding is only raised for https:// URLs, because browsers ignore HSTS sent over plain HTTP. And the clickjacking finding is cleared by either header, so a site that uses only CSP frame-ancestors passes.

Findings are raised for missing headers only. The check shows each value but does not grade it, so a weak CSP or a short HSTS max-age produces no finding. Reading the values is up to you; evaluating HTTP security header policy strength explains how.

What does an example result look like?

The response below is an illustrative example for the documentation domain example.com, not a capture from a real site. It shows the same headers you can print with curl -sI:

HTTP/2 200
content-type: text/html; charset=utf-8
strict-transport-security: max-age=31536000; includeSubDomains
content-security-policy: default-src 'self'; frame-ancestors 'none'
x-content-type-options: nosniff
server: nginx/1.25.3

Read against the checklist, this response would show HSTS, CSP and nosniff as present, and X-Frame-Options as missing with no finding, because frame-ancestors 'none' already covers clickjacking. It would raise three info findings: no Referrer-Policy, no Permissions-Policy, and a Server header that reveals a version.

What does a present Content-Security-Policy tell you?

Content-Security-Policy (CSP) is, per MDN’s CSP documentation, a mechanism to “help detect and mitigate certain types of attacks, including Cross-Site Scripting (XSS) and data injection attacks,” by letting a page restrict which origins scripts, styles, images and other resources may load from. A result showing CSP present is only half the picture — read the value’s directives (default-src, script-src, style-src, and so on). A policy that lists specific, trusted origins is doing real work; one set to default-src * or unsafe-inline for scripts provides little more protection than having no policy at all, because it permits exactly the kind of injected content CSP exists to block. Restricting which origins a script can load from is a separate question from whether that script’s own contents have been tampered with at the origin — Subresource Integrity (SRI) is the complementary check for that, verifying a fetched file’s bytes against a hash the page itself declares.

What does the Strict-Transport-Security row tell you?

A present Strict-Transport-Security header (defined in RFC 6797, which also specifies the max-age and includeSubDomains directives) tells a browser to refuse plain HTTP for the domain going forward, for as long as its max-age specifies — the result shows the full value, including max-age, which matters because a short or absent max-age leaves a wider downgrade window open on repeat visits. See HSTS and HSTS preload explained for the full mechanism, including the “first visit” gap this header alone doesn’t close and what preload adds on top of it.

How do you read X-Content-Type-Options and the clickjacking headers?

X-Content-Type-Options: nosniff is a single-value header with one job: it tells the browser not to guess (“sniff”) a response’s content type from its contents, overriding a declared Content-Type — MIME-sniffing has historically been a vector for getting a browser to execute content as a script that the server never intended to serve as one. MDN is specific about the effect: the browser blocks a style request whose MIME type is not text/css, and a script request whose MIME type is not a JavaScript type. There’s no partial credit here; the result shows it as present with the value nosniff, or missing.

X-Frame-Options and CSP’s frame-ancestors directive both control who can embed the page in an iframe, which is the mechanism behind clickjacking (tricking a user into clicking something they can’t see because it’s layered under a disguised, transparent frame). frame-ancestors is the newer, more flexible control and takes precedence in browsers that support it; a result showing one or the other present is sufficient, and both together is the typical belt-and-suspenders setup for the widest browser compatibility.

Two traps sit behind these rows. Per MDN’s frame-ancestors page, frame-ancestors does not fall back to default-src, so default-src 'none' alone still lets anyone frame the page. And neither control works from an HTML <meta> tag: MDN says X-Frame-Options in a meta element has no effect, and frame-ancestors is not supported there. Both must be real HTTP headers.

What do Referrer-Policy and Permissions-Policy add?

Referrer-Policy controls how much of the current page’s URL gets sent as the Referer header on outbound requests — including to third-party scripts and resources on the same page, not just links a visitor clicks. MDN recommends strict-origin-when-cross-origin as a sensible default, since a missing policy falls back to the browser’s own default, which can leak more of the URL (including query-string contents) to cross-origin requests than most sites intend. The OWASP HTTP Headers Cheat Sheet notes that modern browsers already default to that value, and still advises setting the header explicitly rather than relying on browser defaults.

Permissions-Policy is the newer control for browser features themselves — camera, microphone, geolocation and similar — restricting which origins embedded in the page, such as third-party iframes, may request access to them at all. The OWASP cheat sheet’s example disables three features the page does not need; an empty list () turns a feature off for the page and every frame inside it:

Permissions-Policy: geolocation=(), camera=(), microphone=()

Which headers are not in the checklist, and why?

You may see other security-related headers in a response, or in other tools’ reports. The OWASP cheat sheet gives the current advice on the common ones:

HeaderOWASP adviceWhy it is not a checklist row
X-XSS-ProtectionDo not set it, or set it to 0; use a CSP that disables inline JavaScriptIt is a legacy filter; CSP replaces it
Expect-CTDo not use it; remove it if possibleOWASP notes that Mozilla recommends avoiding it
ServerRemove it or use a non-informative valueCovered by the version-banner finding
X-Powered-ByRemove itDiscloses the framework; not a browser defence
Cross-Origin-Opener-Policysame-origin to isolate the browsing contextAn isolation control beyond the six core headers

How do you read the result as a whole?

The check does not produce one combined score, and reading it that way misses the point: a site missing one header entirely can still be in better shape than one with every header “present” but set to an ineffective value, if the single missing header is a minor one and the “present” policy is permissive. Read each present header’s actual value against the OWASP Secure Headers Project’s reference table, prioritize Content-Security-Policy and Strict-Transport-Security first since they cover the broadest attack surface, and treat nosniff, a clickjacking header and a Referrer-Policy as the next tier. See evaluating HTTP security header policy strength for the specific weak-versus-strong value patterns behind each header in this checklist.

Next step: check your own site

Run AdminKit’s HTTP security headers check against your own URL to see this breakdown for a live response. The check reads one URL per run and does not crawl the site, so test the pages that matter most, such as the login page and any page that embeds third-party content, if they are served differently. If the certificate behind that same site is also worth confirming, how to read a TLS/SSL certificate check result covers the companion check for transport encryption itself, which these headers assume is already in place. See our developers and SaaS teams page for the fuller picture of checking headers alongside a certificate and MCP exposure.

For how this checklist-style result compares to the letter grades and broader reports other free tools produce, see best free HTTP security header analyzers.

This guide is general information, not a security audit of your specific network — see our disclaimer for what AdminKit's guidance does and does not commit to.

Frequently asked questions

Is a 'Present' result the same as 'correctly configured'?

No. 'Present' only means the header was sent in the response at all. A Content-Security-Policy set to default-src *, which allows content from any origin, shows as present but provides close to no protection. Always read the value shown next to a present header, not just the present/missing label — the OWASP Secure Headers Project's reference values are the baseline to compare against.

Which of these headers matters most if I can only fix one?

Content-Security-Policy generally has the biggest practical effect, because a well-scoped policy is one of the strongest available mitigations against cross-site scripting (XSS) — MDN's CSP documentation describes it as letting a site restrict which sources scripts, styles and other resources may load from. Strict-Transport-Security is a close second on any site that still accepts plain HTTP at all, since it closes a protocol-downgrade window that CSP doesn't touch.

What's the difference between X-Frame-Options and CSP's frame-ancestors directive?

Both control whether another site can embed yours in an iframe, which is the mechanism behind clickjacking attacks. X-Frame-Options is the older, simpler header (DENY or SAMEORIGIN); frame-ancestors is a Content-Security-Policy directive that supersedes it with finer-grained control, including allowing a specific list of origins. MDN notes browsers that support frame-ancestors ignore X-Frame-Options when both are present, so a site can send frame-ancestors alone on modern clients, keeping X-Frame-Options only as a fallback for older ones.

Why does the result flag a missing Referrer-Policy as an issue, when my site has no sensitive URLs?

The referrer leak isn't about your site's own URLs being sensitive — it's that, without a Referrer-Policy, a browser's default behavior still sends the full referring URL (which can include session tokens, search terms or internal path structure in the query string) to every external link a visitor clicks, including third-party scripts and images. MDN recommends strict-origin-when-cross-origin as a safe default: it sends the full URL only to same-origin requests and just the origin elsewhere.

Does this check mean my site has no other vulnerabilities?

No — it is a configuration checklist for one category of browser-enforced defenses, not a full security audit. A site can send every recommended header correctly and still have an unrelated vulnerability in its application code, authentication, or server configuration that headers alone can't prevent.

Want AdminKit to watch this for you?

Free checks tell you the state today. Protective DNS and monitoring tell you when it changes.

Huy Nguyen (David), Founder and CEO of Netbase JSC

Talk to an expert

Huy Nguyen (David), Founder & CEO, Netbase JSC

Ask whether this fix applies to your setup, or get a second pair of eyes on your DNS and mail records.