Security misconfiguration is the OWASP Top 10 category for software that runs perfectly and is still insecure, because it was left in a default, verbose or over-permissive state. It is not a coding bug. It is a setting someone never changed, a file someone never deleted, or a header someone never added — and it is one of the easiest findings to confirm from the outside.

The good news is that almost every item in this category has a fix you can apply in minutes, without touching application code. The hard part is knowing what to look for. This article lists ten concrete security misconfiguration examples, with a one-line test and the corresponding fix for each.

Why this matters for your website

Misconfigurations are attractive to attackers because they scale. Nobody targets your shop by name; they scan whole IP ranges for /.env, for Index of /, for admin panels answering with factory credentials. Your site is found by pattern, not by reputation.

They also sit upstream of other breaches. An exposed .env file hands over your database password and API keys in one request, which is a credential leak, a data breach and a supply-chain problem at once. OWASP has ranked Security Misconfiguration in its list for years — it was A05:2021 in the 2021 edition — and it remains in the current OWASP Top 10 2025 set of risks. See the OWASP Top 10 project for the canonical definitions.

Ten security misconfiguration examples, tested and fixed

Each test below is safe to run against your own site from a terminal or a private browser window.

  • Directory listing. Test: open https://example.com/uploads/ and look for "Index of /". Fix: Options -Indexes in Apache, autoindex off; in nginx.
  • Verbose errors and stack traces. Test: request a broken URL such as /?id=' and check whether file paths or SQL appear. Fix: set display_errors=Off, APP_DEBUG=false, and serve custom 404/500 pages.
  • Default credentials. Test: try the vendor's documented defaults on every admin panel, router and database UI you own. Fix: change them, enable multi-factor authentication, restrict the panel by IP or VPN.
  • phpinfo and debug endpoints. Test: request /phpinfo.php, /info.php, /debug and /_profiler. Fix: delete the files and strip debug toolbars from production builds.
  • Open cloud storage buckets. Test: open the bucket or container URL in a logged-out browser. Fix: enable block-public-access at account level and serve private objects through signed URLs or a CDN.
  • Missing security headers. Test: curl -sI https://example.com. Fix: add Content-Security-Policy, HSTS, X-Content-Type-Options, Referrer-Policy, Permissions-Policy and X-Frame-Options — see our guide to HTTP security headers.
  • Permissive CORS. Test: send Origin: https://evil.example and see whether the server reflects it back. Fix: allowlist exact origins, and never combine a wildcard with Access-Control-Allow-Credentials: true.
  • Unnecessary HTTP methods. Test: curl -X OPTIONS -i https://example.com and read the Allow header for PUT, DELETE or TRACE. Fix: limit methods to GET, POST and HEAD at the web server or WAF.
  • Default sample applications. Test: request paths such as /examples/, /manager/html, /phpmyadmin/ or a CMS's bundled demo plugin. Fix: uninstall them; do not merely password-protect.
  • Exposed backup and config files. Test: request /.env, /.git/config, /wp-config.php.bak and /backup.zip. Fix: move them out of the webroot, deny dotfiles in the server config, and rotate every secret that was reachable.
curl -sI https://example.com | grep -iE 'content-security|strict-transport|x-frame'
curl -s -o /dev/null -w '%{http_code}\n' https://example.com/.env
curl -X OPTIONS -i -s https://example.com | grep -i '^allow'

If the .env check returns 200 rather than 404 or 403, treat every credential in that file as compromised and rotate it before anything else.

Two-column diagram comparing a leaky HTTP response with a hardened one
The same request to the same site: a misconfigured server volunteers paths and versions, a hardened one gives nothing away.

A misconfiguration checklist by layer

Misconfigurations cluster by layer, and so should your review. Work down this list once per quarter and after every migration.

  • Web server (Apache, nginx, IIS). Directory listing off; server version tokens suppressed; dotfiles denied; HTTP methods limited; TLS 1.2+ only with a valid, monitored certificate.
  • Framework and application. Debug mode off; unique application secret keys per environment; custom error pages; CORS allowlist; cookies set with Secure, HttpOnly and SameSite.
  • CMS (WordPress, WooCommerce, Magento, Drupal). Unused plugins and themes deleted, not just deactivated; file editing disabled (DISALLOW_FILE_EDIT); XML-RPC closed if unused; admin accounts on unique names with MFA; automatic security updates on.
  • Cloud and infrastructure. Public access blocked on storage buckets; least-privilege IAM roles; no 0.0.0.0/0 rules on SSH, RDP or database ports; logging and alerting enabled; staging environments behind authentication and noindex.

Staging deserves special attention. A copy of production with debug on, sample data and no password is a full disclosure of your stack.

How to check your own site

Start free and unauthenticated. You can check your security headers in seconds, and the Website Security Check gives a quick external read on TLS and common exposures. Open-source tooling goes further: nuclei ships thousands of misconfiguration templates, nikto probes for default files and dangerous methods, and OWASP ZAP passively flags header and cookie issues while you browse.

Scantinel (scantinel.ai) is an automated website security assessment platform that runs OSINT, network, and OWASP Top 10 web vulnerability tests against ownership-verified assets and delivers actionable, business-owner-legible reports. The Essential scan is passive and finishes in minutes: security headers, cookie flags, TLS posture, DNS and email records, plus exposed common files such as .env, .git and backup archives. The Full scan adds active testing with nuclei, nikto and OWASP ZAP, a port scan with service fingerprinting, and a CMS deep scan via WPScan for plugin-level issues.

Be clear about the limits. An automated assessment is not a penetration test, and it will not reason about your business logic the way a human reviewer does. It is very good at the repetitive external checks that misconfigurations fall into. If you would rather not run the tooling yourself, you can scan for misconfigurations against assets you have verified you own.

Key takeaways

  • Security misconfiguration is a settings problem, not a code problem — which means most fixes are fast and free.
  • The highest-impact checks are exposed .env and .git files, default admin credentials and public cloud storage.
  • Missing security headers are real findings: they remove a defence layer even when nothing is exploitable today.
  • Review by layer — server, framework, CMS, cloud — after every migration, plugin install or environment clone.
  • Rotate any secret that was publicly reachable, even briefly; deleting the file is not enough.

Frequently asked questions

What is an example of security misconfiguration?

A classic example is a .env file left in the webroot, which returns database credentials and API keys to anyone who requests it. Other common cases are directory listing left enabled, a stack trace shown on an error page, or an admin panel still using the vendor's default password.

Why is security misconfiguration so common?

Defaults are chosen for ease of installation, not for production safety, and nothing visibly breaks when they stay in place. Sites also drift: a migration restores an old config, a developer enables debug mode to chase a bug, a plugin adds an endpoint. Without a recurring external check, the drift is invisible.

How do I find misconfigurations on my website?

Begin with the one-line tests above, then run a free headers check and a nuclei or nikto pass against your own domain. For continuous coverage, an automated assessment against ownership-verified assets will repeat those checks on a schedule and flag what changed.

Published February 2025 by the Scantinel Security Research Team. Verify current guidance against primary sources — OWASP, NVD and WPScan — before acting on any finding.