security-txt.eu

GuideAs of 9 min read

"security.txt not configured": what the finding means and how to close it

A scanner, a penetration test report, Cloudflare or a customer's security questionnaire says your security.txt is missing, not found or not configured. As far as attack surface goes, the finding is harmless, and in most cases you can close it in fifteen minutes. Sometimes, though, it appears even though the file has been there for months. This page covers both cases and shows how to prove the fix.

What the finding means

security.txt is a small text file at https://your-domain/.well-known/security.txt. It tells security researchers, CERTs and automated tools where to report a vulnerability they have found. The format is defined in RFC 9116; the guide explains how it is built.

Strictly speaking, "security.txt missing" means: when the tool requested that URL, it did not get back anything it recognised as a valid security.txt. That is not the same as "you don't have one". The scanner may have queried a different host, stopped at a redirect or a bot challenge, or received an error page instead of the file. Before changing anything, find out what the scanner actually saw.

How serious is it?

A missing security.txt is not a way into your systems. Scanners therefore list it as informational, not as a vulnerability: Tenable, for example, rates its check "Info". If someone presents the finding as critical, ask why.

It still matters, for a different reason. The file is the reporting channel for the day someone outside finds a real flaw. Without it, that report ends up in a general inbox, a sales contact form or nowhere at all, and some finders then go public instead. Three reasons to close the finding rather than argue it away:

  • Little effort. One file with two required fields, uploaded once and renewed once a year.
  • Manufacturers under the Cyber Resilience Act. From 11 December 2027, anyone placing products with digital elements on the EU market must provide a contact address for reporting vulnerabilities (Regulation (EU) 2024/2847, Annex I Part II point 6) and include evidence of it in the technical documentation (Annex VII point 2(b)). The CRA does not prescribe a format, but security.txt is the channel that tools discover automatically. Whether this applies to you is covered in Is security.txt required?
  • BSI TR-03183-3. If you follow the German BSI's technical guideline, you must make sure security.txt can be found automatically, explicitly including crawlers such as internet.nl, and adjust firewall and DDoS rules accordingly (section 4.2.11 a). Under the TR, a scanner that cannot find your file is a real gap.

If the finding comes from a customer's vendor questionnaire or a cyber insurance assessment, the file plus a check result usually settles it in one line. How to provide that evidence is described further down.

Which tools report it

The wording varies by tool. Some examples we were able to verify:

ToolFindingRatingWhat is checked
Cloudflare Security Insights"Security.txt not configured" on the domain overviewMissing configurationWhether Cloudflare's own security.txt feature is set up (see Cloudflare below)
Tenable Web App Scanning"Security.txt File Not Detected" (plugin 112723)InfoWhether a security.txt file is detected on the target
internet.nlPart of the website testRecommended standard; according to the 2022 announcement not counted in the overall scoreWhether the file exists and is valid under RFC 9116
Nuclei (template security-txt)"security.txt File" as a finding when the file is presentinfoRequests /.well-known/security.txt, then /security.txt; matches on status 200 and "Contact:" or "Expires:" in the body

Nuclei works the other way round: the template reports the file when it finds it. If that finding is absent from the report, testers conclude the file is missing. The template follows at most two redirects and only matches files up to 4,096 bytes. A normal security.txt is far smaller, even when signed, but keep an eye on size if yours carries a lot of comments or fields.

Vendor risk rating platforms, external attack surface scans and pentest reports list the finding too, often under headings such as "Missing security.txt" or "Vulnerability disclosure contact not found". How much weight it carries there is up to each vendor and rarely published.

Why it shows up even though you have a file

If you do have a security.txt and the scanner still says it is missing, the cause is almost always one of these:

CauseHow to spot itFix
Wrong pathThe file is only at /security.txtMove it to /.well-known/security.txt; the old location may redirect (RFC 9116 section 3)
Different host scannedThe scanner queries company.eu while the file only exists on www.company.eu, or it checks a subdomain such as shop.company.euServe the file on every host that gets scanned; it only applies to the host it is fetched from (RFC 9116 section 3.1)
Redirectcurl -sIL shows a 301 or 302, for example to the home page or another domainServe the file directly; redirects to other domains make tools suspicious, and under the TR the canonical URI must be reachable without redirects (4.2.2 c)
Bot protection or challengeFine in the browser, but curl gets 403, 429 or 503, or an HTML challenge pageExempt /.well-known/ from bot protection; Cloudflare notes are in the guide
Error page with status 200The response starts with <!DOCTYPE html>The CMS intercepts the path: create a real file or a rule in the CMS, see security.txt on your platform
Wrong content typeThe header shows text/html or application/octet-streamConfigure the server to send text/plain; charset=utf-8 (RFC 9116 section 3)
HTTP onlyThe https:// request fails, http:// worksThe file must be served over HTTPS (RFC 9116 section 3)
Expired or missing required fieldsExpires is in the past or absent, Contact is missingGenerate a new file; tools that validate the format will otherwise treat it as invalid. Details in security.txt expired

Diagnose it yourself with curl

Three commands show what a scanner sees. Use the host named in the finding, not the one you think is right:

# 1. Status, content type and redirects
curl -sIL https://company.eu/.well-known/security.txt

# 2. The first lines of the body
curl -sL https://company.eu/.well-known/security.txt | head -n 15

# 3. Like a simple tool without a browser user agent
curl -sI -A "security-txt-check" https://company.eu/.well-known/security.txt

You are in good shape if the first command ends directly in HTTP/2 200 with content-type: text/plain; charset=utf-8, the second starts with lines such as Contact: and Expires: (preceded by -----BEGIN PGP SIGNED MESSAGE----- for a signed file), and the third returns the same as the first. If the third gets a 403 or an HTML page, a bot filter is blocking exactly the tools the file is meant for.

Repeat the commands with and without www and for every subdomain named in the finding. The checker is quicker: it fetches the file from outside, shows the redirect chain and flags bot protection, HTML instead of text, wrong charset and expired dates, each with the relevant RFC section.

Fix it in fifteen minutes

If there is no file yet, this is the shortest route:

  1. Pick the address. A role mailbox such as security@company.eu that at least two people read. Not info@ and not a personal address. The reasons are in Setting up a security contact.
  2. Generate the file. Enter domain and address in the generator and choose the standard (RFC 9116 for an ordinary company website, BSI TR-03183-3 for manufacturers). It sets a valid expiry date and only releases the file once it is error-free.
  3. Upload. Save it as security.txt in the .well-known folder at the web root. The folder name starts with a dot, so some FTP clients hide it. WordPress, nginx, IIS, Shopify and other systems have their own routes, see security.txt on your platform.
  4. Check. Run the checker for every host in the finding, with and without www.
  5. Send a test email. Mail the address in the file and see whether it arrives and who reads it. No scanner tests this, but it is the whole point.
  6. Set a reminder. Put the expiry date in the calendar with some lead time, or the finding comes back next year as "expired".

A file that satisfies most scanning services looks like this:

Contact: mailto:security@company.eu
Expires: 2027-10-26T23:00:00.000Z
Preferred-Languages: en, de
Canonical: https://www.company.eu/.well-known/security.txt
Policy: https://www.company.eu/security-policy/

RFC 9116 only requires Contact and Expires. What the other fields do and which ones the TR adds is explained in All security.txt fields; ready-made templates for both standards are among the examples.

Cloudflare: "Security.txt not configured"

If your domain runs through Cloudflare, the notice appears on the domain overview under Security Insights. Since September 2024 Cloudflare has offered its own security.txt feature, found under Security, Settings, filter "Web application exploits", section Security.txt. There you enter Contact and Expires, and optionally Encryption, Acknowledgments, Canonical, Hiring, Policy and Preferred-Languages.

That leaves two options:

  • Use Cloudflare's feature. Convenient if you have no access to the server. The form has no option for a signature, and no CSAF or Bug-Bounty fields. For a file under BSI TR-03183-3, which must be signed (4.2.10), that is not enough.
  • Keep your own file. If your server already serves a security.txt, use the checker to confirm it gets through Cloudflare without a bot challenge. If it does, you can archive the notice in Security Insights. Cloudflare explicitly provides for archiving an insight once you have understood it and decided not to act on it.

With Cloudflare, bot protection matters more than the notice itself: it is the most common reason other scanners cannot find a file that shows up fine in the browser.

Provide evidence and get a rescan

Once the file is live, your customer, insurer or scanning service will usually want proof. Three pieces work well:

  • The URL of the file, for example https://www.company.eu/.well-known/security.txt. Anyone can open it.
  • A dated check result, such as a screenshot or PDF from the checker showing status, content type, expiry date and, if present, a valid signature.
  • One sentence on process: who reads the mailbox and how quickly you respond. In questionnaires, that is often the real question behind the finding.

Then have the finding rechecked. On internet.nl, simply run the test again. With scanners such as Tenable, the next scheduled scan or a manual rerun is enough. Rating platforms refresh their data on their own schedule; most offer a way in the customer portal to mark a finding as resolved or to dispute it. After a pentest, include the curl output or the check result in the retest.

To keep it from coming back: most "missing" findings in year two are caused by an expired Expires date, a website migration or a new plugin that intercepts the path.

Monitoring fetches the file from outside every day, just like a scanner, and alerts you if it disappears, becomes unreachable or is about to expire. 60 euros net per domain per year.

Frequently asked questions

What does "security.txt not configured" mean?

A tool did not find a valid security.txt at /.well-known/security.txt. In Cloudflare, the notice refers to Cloudflare's own security.txt feature in the dashboard. Either way, fetch the file with the checker to see whether it is really missing or just unreachable.

Is a missing security.txt a vulnerability?

No. Scanners rate it as informational; Tenable, for example, uses severity "Info". It shows that a standard reporting channel for vulnerabilities is missing, not that your website can be attacked.

My security.txt opens in the browser. Why does the scanner say it is missing?

Usually bot protection blocks automated requests, the scanner checks a different host (without www, or a subdomain), or it does not follow a redirect. The curl commands above and the checker show which one it is.

Is /security.txt in the web root enough?

No. The location that counts is /.well-known/security.txt (RFC 9116 section 3). The root location is only allowed for legacy compatibility and may redirect to the new one.

Do I need security.txt on every subdomain?

The file only applies to the host it is fetched from (RFC 9116 section 3.1). If a scanner checks a subdomain with its own website, that subdomain needs its own file or a redirect to the main domain's file.

How long does it take to fix?

With the generator and access to the web server, about fifteen minutes. It takes longer if you first have to create a role mailbox or an exception in your bot protection.

What should I send a customer or insurer as proof?

The URL of the file, a dated result from the checker and a sentence on who reads the mailbox. That also answers the question about your reporting channel, which is usually what the finding is really about.

Sources