What does the tool do?

The tool parses the one pasted response header block and reviews six headers: Strict-Transport-Security, Content-Security-Policy, Content-Security-Policy-Report-Only, X-Frame-Options, Referrer-Policy and X-Content-Type-Options. Each one is shown as “missing from this block”, “present” or “present, review notes”, with short source-based notes. No request is made to any address; this is not a scanner, scorer or certificate tool.

The output is a lexical reading of the pasted text. It cannot prove that the response arrived over HTTPS, skipped caches and proxies, or matches what the server sends now. Findings are points to review, not a verdict of “secure” or “insecure”.

How do you get the header block?

Open DevTools, select the document request in the Network tab and copy the lines under Response Headers. If the page redirects, choose the final response, not an intermediate one. In your own terminal you can use a command like this synthetic example:

curl -sS -D - -o /dev/null https://example.test/

Following redirects with -L prints several response blocks; the tool rejects that and asks you to pick one. HEAD and GET can return different headers. Remove cookie and authorization headers whenever you share a block with other people; this tool neither keeps nor shows them.

How is each header read?

  • HSTS: max-age, includeSubDomains, preload, duplicates and syntax are read. Not claimed: that the response came over HTTPS or that subdomains are ready.
  • CSP: directives, duplicates, 'unsafe-inline', 'unsafe-eval', *, data:, nonce and hash shape, and frame-ancestors are read. Not claimed: how a browser judges every resource.
  • Report-Only: the same analysis in a separate card, plus the reporting directive. Not claimed: that reports reach a server.
  • X-Frame-Options: DENY, SAMEORIGIN, obsolete ALLOW-FROM and multiple values are read. Not claimed: ALLOW-FROM support.
  • Referrer-Policy: the token list is read; the last recognized token counts. Not claimed: the browser default.
  • X-Content-Type-Options: the first value is checked for nosniff. Not claimed: that a missing header is exploitable.

The “Not claimed” part of each item is the verdict the tool deliberately withholds: a missing X-Content-Type-Options is noted as a missing hardening header, not a vulnerability.

CSP: enforced, Report-Only and several policies

Content-Security-Policy enforces a policy, while Content-Security-Policy-Report-Only only reports and blocks no resource. The tool shows them in separate cards and flags a block that has only Report-Only. Without report-to or report-uri, a Report-Only policy sends nothing to a server, and a reporting directive does not prove that reports arrive.

Several CSP headers, or several comma-separated policies, are enforced together: a request must pass every policy, and a later policy cannot loosen an earlier one. The tool neither merges them nor lets the last one win; it numbers each policy. default-src is not a fallback for every directive, so frame-ancestors, base-uri and form-action need their own entries, and the tool notes those gaps. It does not claim to resolve nonce and 'strict-dynamic' interaction like a full browser evaluator.

Privacy and masking

The text is processed in your browser and never written to a server, share link, address bar, local storage or log. A share link carries only the delivery option. The text box is masked by default; clear removes text and result.

Values of Cookie, Set-Cookie, Authorization, Proxy-Authorization, X-Api-Key and X-Auth-Token are dropped during parsing, whatever their capitalization. Unknown headers are only counted, never echoed. Error messages contain a line number and a reason, never the raw line. The copy button copies only this masked summary.

Limits

The tool expects one response block: at most 64 KiB, 200 headers and 8192 characters per line. Several responses, a redirect chain, a body, line folding, control characters or an empty name produce an explicit error; the tool never silently picks one. HSTS applies only to HTTPS responses, which a block cannot prove, so the default delivery option is “unknown”.

Caching and CORS headers are out of scope and receive no definitive verdict. The tool produces no score, certificate or “secure” result.

Reading the example

The “Load example” button inserts a synthetic response with an example.test host and an invented cookie. It yields real findings: 'unsafe-inline' in script-src, * in img-src, the obsolete ALLOW-FROM, and a separate Report-Only policy. The cookie value stays hidden; only a “Set-Cookie ×1” note appears.

The hash generator computes digests, but a CSP source expression still needs the 'sha256- prefix and single quotes around it. The SPF, DKIM and DMARC checker is a similar local analysis for email records; the .htaccess redirect generator covers redirect rules, not security headers.

Sources and last check

Last checked: 25 October 2026. The primary sources below were opened on that date and the relevant sections checked.

Frequently asked questions

Does this tool scan my site?

No. It reads only the text you paste and requests no address. There is no URL field, crawler or proxy, so the result describes your block, not the current state of the site.

Does a Report-Only policy protect anything?

No. It reports violations but blocks no resource; only the Content-Security-Policy header is enforced. The tool shows both separately and warns when only Report-Only exists.

Does a missing header mean there is a hole?

No. A missing header means it does not appear in this block. Another layer may provide the same protection, or the header may not be needed.

Does the share link contain my headers?

No. The link carries only the delivery option. Pasted text never enters the address bar, storage or a network request.

Why can I not see cookie values?

Values of Cookie, Set-Cookie, Authorization, Proxy-Authorization, X-Api-Key and X-Auth-Token are dropped during parsing and appear in no message or summary; you see only the name and line count.

Published: · Updated: