Cybersecurity

Web application security: the flaws to fix first

Web application security in plain language: the flaws attackers exploit most, how to check your own app, and what to demand from your developers.

Your business runs on a web application: a client portal, an internal tool, an online ordering system. It holds customer data, invoices and documents, and it is reachable from anywhere. You are not a security specialist, and you do not need to be. But you do need to know which questions to ask.

Web application security is less mysterious than it sounds. Most serious problems belong to a handful of well-known families of flaws. In this guide we explain them in plain language, show you how to check where your own application stands, and give you a checklist of what to demand from the people who build and maintain it.

Web application security starts with the OWASP Top 10

The Open Worldwide Application Security Project (OWASP) is a non-profit foundation that publishes free security guidance. Its best-known document, the OWASP Top 10, ranks the most critical risks for web applications. It is revised every few years; as of October 2026, the current edition is the OWASP Top 10:2025.

The Top 10 is not a full security standard. It is a shared vocabulary: when you ask a developer “how do you handle broken access control?”, you both know what you mean. That makes it the right starting point for an owner or an IT manager.

From the ten categories, we single out six families that we would check first on almost any business application.

The six families of web application vulnerabilities to fix first

1. Broken access control

The application lets a user see or do something they should not. A customer changes a number in the address bar and opens another customer’s invoice. An employee reaches an admin page simply by typing its address. A file link keeps working for anyone who has it, without a login.

Access control ranks first in the 2025 edition. The fix is a rule on the server for every page, every action and every file: who may do what, on which record. Hiding a button is not access control.

2. Injection, including SQL injection

An injection happens when the application takes what a user types and passes it, unchecked, to another system: the database, the operating system, a document template. With SQL injection, a crafted entry in a form or a web address changes the database query itself, and can read or delete data.

The fix is well known: parameterised queries (the data is always sent separately from the command), validation of every input, and escaping of every output. Injection is no longer at the top of the list, but it still turns up in older applications built by hand.

3. Weak authentication

Passwords stored in a readable form, no limit on login attempts, no second factor for administrators, sessions that never expire, a “remember me” cookie that holds the password itself. Each one makes an account easier to take over.

What to expect today: passwords stored with a modern hashing function, a lock or delay after repeated failures, two-step login at least for staff and admins, and sessions that end.

4. Outdated components

A web application relies on a language, a framework and dozens of libraries. Each one receives security fixes for a limited time. An application still running on a version that no longer gets updates, what is called end-of-life software, carries every flaw discovered since, with no fix coming. The critical PHP vulnerabilities of October 2026 are a recent example.

This is often a symptom of a wider problem, which we cover in our article on technical debt and modernisation.

5. Security misconfiguration

The code may be sound while the server is not: debug mode left on in production, error pages that show file paths and database details, default admin accounts, an upload folder that accepts any file type, storage open to the internet. Misconfiguration ranks second in the 2025 edition, because it is easy to get wrong and easy to miss.

6. Missing logs and alerts

If nobody records who logged in, who changed a status or who downloaded a file, you cannot tell whether something went wrong, or when. Logs should capture the sensitive events, be kept somewhere an attacker cannot easily erase them, and trigger an alert on what really matters, such as a burst of failed logins on an admin account.

How to check your own application

  1. List what the application holds. Personal data, financial data, documents, credentials for other systems. That tells you how much protection it needs.
  2. List who can access it. Roles, external users, former staff, shared accounts. Shared accounts and accounts of people who have left are a frequent and easy fix.
  3. Ask for the component inventory. Language and framework versions, main libraries, and the date each stops receiving security updates.
  4. Look at the basics yourself. Is every page served over HTTPS? Does an error show technical details? Does a document link still open after you log out?
  5. Read the last security test, if there is one. When was it done, what was found, and what was actually fixed?
  6. Then commission an independent review for anything holding sensitive data: a code review, a penetration test (an authorised, simulated attack), or both.

Never test an application you do not own, or run attack tools without written permission from the owner and the host. If in doubt, leave the testing to a professional.

What to demand from your developers: a review checklist

Whether the team is in-house or external, these are reasonable expectations for a business application. Use the list in a meeting, or attach it to a request for proposal. If you sell your application to large companies, expect their security questionnaires to go further, with single sign-on, audit logs and backups: our security checklist for selling SaaS to large companies covers what they ask.

Area What to ask for A good answer looks like
Access control How is each page, action and file protected? Server-side rules per role and per record, with automated tests
Injection How are database queries built? Parameterised queries everywhere; inputs validated, outputs escaped
Authentication How are passwords and sessions handled? Modern hashing, attempt limits, two-step login for staff, session expiry
Components Which versions, and until when are they supported? An up-to-date inventory and a regular update routine
Configuration How do production settings differ from development? Debug off, generic error pages, no default accounts, uploads restricted by type
Logs What is recorded, and who is alerted? Logins, permission changes and downloads logged; alerts on anomalies
Process How is security checked before each release? Automated checks in the delivery pipeline, plus code review
Secrets Where are passwords and keys stored? Outside the code, never in the version history

That last “Process” line matters more than it seems. At Orange Business Services, in a project led by our founder, Anass Chabbab, in his previous roles, continuous integration with Jenkins and SonarQube checked the code automatically on every change. Read the Orange Business Services case. The same idea applies to security: checks that run on every release catch problems before they reach production.

If your provider cannot answer most of these questions clearly, that is a signal in itself. Our guide to choosing an IT provider covers the other criteria. And if the weak point is the wider organisation rather than one application, start with our article on cybersecurity for SMEs.

When a fix is not worth it

Not every finding deserves a project. A low-risk internal tool used by three people may only need tighter access and updated components. But when an application stacks several of these flaws on an unsupported platform, rebuilding can cost less than patching: a decision to make with figures, not in a hurry.

Frequently asked questions

What is the OWASP Top 10?

It is a list of the ten most critical security risks for web applications, published free of charge by the OWASP foundation and revised every few years. As of October 2026, the current edition is the 2025 one. It is not a certification and does not cover everything, but it is the shared reference that developers, testers and buyers use to talk about web application vulnerabilities.

Is SQL injection still a risk today?

Yes, mainly in older applications or in code written without a modern framework. Current frameworks make parameterised queries the default, which removes most of the risk when they are used properly. The danger comes from hand-built queries that insert user input directly. A code review can find them quickly, and fixing them is usually straightforward.

How often should a web application be tested?

There is no single rule. A sensible rhythm is an automated check on every release, a review of components every month or so, and an independent test once a year or after any major change, such as a new module, a new type of user or a move to new hosting. Applications holding sensitive data justify a closer rhythm.

In short

Most serious flaws in business web applications fall into a few families: broken access control, injection, weak authentication, outdated components, misconfiguration and missing logs. The OWASP Top 10 gives you the vocabulary to ask about them. Start by listing what your application holds, who can reach it and what it depends on, then ask your developers for clear answers and an independent test where the data is sensitive.

This article gives practical pointers, as of October 2026. It does not replace an assessment by a qualified security professional, and any legal obligation must be confirmed by a lawyer in your country.

Not sure where your application stands? Book a free 30-minute discovery call: we look at it with you and tell you which flaws to fix first. You can also see how we build and maintain web applications.

Related servicesWebsites and web appsCustom softwareModernisation

Have a project in mind?

Let’s talk about it on a 30-minute discovery call, free and with no obligation.

Book a discovery call

Further reading

All articles