Logging in is authentication. Authorization is a separate question: whether this logged-in user may open a particular page or record. Confusing the two is the source of many serious vulnerabilities, including IDOR, so the permission check has to run on every request.
-
Authentication and authorization
The first checks who you are, the second what you are allowed to do.
-
Backup
A copy of the files and database kept outside the website’s server, which lets you restore the site after a failure or an attack.
A copy stored only on the same server as the website is not a backup. In a serious failure it disappears together with the original. Every so often, check that restoring really works.
In practice: Dyżur: a daily database backup and a copy you can download with one click
-
Brute-force attack
Guessing a password by automatically trying thousands of combinations.
On WordPress sites it goes on constantly: bots try the username admin and popular passwords. The effective defences are long, unique passwords, a limit on failed login attempts and two-factor authentication. Changing the login address only helps as an extra.
-
CSP (Content Security Policy)
A header that tells the browser where the website may load scripts, styles and images from.
If someone injects a malicious script (XSS), a well-configured CSP can block it, because the script doesn't come from an allowed source. It is rolled out carefully: first in report-only mode to see what it would block, and only then is blocking switched on.
In practice: Guide: how to secure WordPress, including security headers
-
CSRF
An attack in which a foreign website forces a logged-in user to perform an action they didn’t intend.
A logged-in administrator opens a planted page, and it sends a request on their behalf, for example to change the password. One-time tokens in forms (nonces in WordPress) and the SameSite cookie attribute protect against this. Many plugins have vulnerabilities because they lack these safeguards.
-
CVE
The number of a publicly known vulnerability in a specific piece of software, for example CVE-2021-41773.
If a vulnerability has a CVE number, there is usually a fix too, most often in a newer version of the software. The number makes it easy to check whether the vulnerability affects your version.
In practice: Rysa: a scan result matched against CVE numbers
-
CVSS
A score from 0 to 10 of how dangerous a vulnerability is in itself. From 7.0 it is high, from 9.0 critical.
The number knows nothing about your business. Urgency also depends on whether the vulnerability is reachable from the internet and whether anyone is already exploiting it.
In practice: Guide: how to read a security audit report
-
CWE
A catalogue of types of software weaknesses, with numbers, for example CWE-79 for XSS.
A CVE describes a specific vulnerability in a specific product, while a CWE describes the class of flaw that such vulnerabilities come from. Audit reports often give both numbers, so the developer knows which error pattern to fix, not just one spot in the code.
-
GDPR
The EU regulation on personal data protection, called RODO in Poland and enforced there by UODO (the Polish data protection authority); on a website it covers contact forms and cookies, among other things.
The minimum is a privacy policy describing who processes the data and for what purpose, and collecting only the data that is needed for that purpose.
-
HSTS
A header that tells the browser to connect to the website only over HTTPS.
After the first visit, the browser switches http to https by itself, so an attacker on a public network can't serve an unencrypted version. Before turning it on you have to be sure the whole website and all subdomains work over HTTPS, because undoing it is slow.
In practice: Guide: how to secure WordPress, including security headers
-
IDOR
A flaw that lets you see someone else’s data by changing a number in the address, for example invoice/101 to invoice/102.
The application checks that you are logged in, but not whether the resource is yours. It is one of the most common flaws in web applications and a frequent finding in audits. It is fixed with a server-side permission check on every read.
-
OWASP Top 10
A list of the ten most common categories of web application vulnerabilities, maintained by the OWASP organisation.
It is the reference point for audits: instead of looking for “anything at all”, you check the best-documented classes of flaws one by one.
In practice: Guide: OWASP Top 10 without jargon
-
Password hashing
Storing passwords as an irreversible hash instead of the password itself.
If the database leaks, the attacker gets hashes, not passwords. This is done with slow algorithms designed for passwords (for example Argon2, bcrypt), not plain MD5 or SHA-1, which can be cracked at billions of attempts per second. A password should never be visible in code or in logs.
-
Penetration testing (pentest)
A controlled attempt to break into an application, carried out with the owner’s consent, to find weaknesses before an attacker does.
It ends with a report: what was found, how serious it is and in what order to fix it. After the fixes, a retest is done.
In practice: Rysa: a phone as a station for checking networks
-
Rate limiting
A limit on how many times in a given period one user can trigger a given action.
It protects login and contact forms from password guessing and spam flooding, and APIs from overload. The limit is applied on the server, not in the browser, because the attacker controls the browser.
-
security.txt
A file on the website that tells researchers how to report a vulnerability they have found.
It lives at /.well-known/security.txt and contains a contact, the language for reports and an expiry date. Thanks to it, a person who found a flaw knows whom to write to, instead of publishing it or giving up. It is described by the RFC 9116 standard.
-
SQL injection
An attack that injects your own query into the database through an ordinary form field.
A successful attack can read or delete data. The defence is simple and has been known for years: parameterised queries instead of gluing text together.
-
SSL / HTTPS
Encryption of the connection between the browser and the server. It is where the padlock next to the site’s address comes from.
Without HTTPS, browsers mark the site as unsafe and data from forms travels as plain text. The certificate is free today and renews automatically.
-
SSL/TLS certificate
A certificate that lets the browser encrypt the connection and confirm that this really is that website.
Thanks to it, the address starts with https and the browser shows a padlock. Free certificates (for example Let's Encrypt) are fully sufficient for a typical business website and renew automatically. An expired certificate shows visitors a warning instead of the site.
-
SSRF
An attack in which a server is forced to send a request to a place the attacker can’t reach themselves.
If an application fetches an address supplied by the user (for example a link preview), the attacker can point it at an internal address, not reachable from the internet. The defence is an allowlist of permitted addresses and blocking internal networks, not just checking whether the address looks valid.
-
Supply chain attack
A break-in through someone else’s code that you trust: a plugin, a library or a developer tool.
Instead of attacking your website, the attacker poisons a popular plugin or package, and the update delivers them access. It is limited by a minimum of dependencies, checking who maintains an add-on, updating after a short wait and backups you can quickly roll back to.
-
Two-factor authentication (2FA)
Logging in with a password plus a second proof, for example a code from an app or a hardware key.
Even if the password leaks, nobody can log in without the second factor. Hardware keys and code-generating apps are best; SMS codes are convenient but weaker. For a website administrator's account, it is one of the cheapest and most effective safeguards.
In practice: Guide: how to secure WordPress, including two-factor login
-
WAF (Web Application Firewall)
A filter in front of the website that rejects requests that look like an attack.
It checks traffic for known patterns, for example SQL injection attempts, before it reaches the application. It is an extra layer: it doesn't fix a vulnerability in the code, it only makes it harder to exploit. Rules that are too aggressive can block real users, so they need tuning.
-
XSS
Cross-Site Scripting: injecting a foreign script that then runs in another user’s browser.
It is used, among other things, to hijack the sessions of logged-in people. The protection is consistent encoding of any data displayed on the page.
-
Zero-day
A vulnerability that attackers already know about and for which there is not yet a fix.
The vendor has zero days to fix it, hence the name. Updating alone doesn't protect against a zero-day, which is why layered defence matters: backups, a minimum of plugins, limited permissions and monitoring that notices unusual traffic.
Ready to start?
Let’s start with a conversation
Describe what you want to build. The reply contains the scope, the price and the timeline. The inquiry costs nothing and commits you to nothing.