
In short
- The OWASP Top 10 is a widely used list of the most serious web application security risks. The current edition is 2025.
- Many of the risks can be checked with ordinary testing skills: changing an id in a URL, sending unexpected input, logging in and out.
- Testers don’t replace security specialists, but they catch many of these problems earlier and more often.
On this page
A tester is checking the order history page. Out of curiosity, they change the number at the end of the URL from 10432 to 10431. Someone else’s order appears: name, address, what they bought. No security scanner flagged it, because to a scanner the page worked perfectly. It took a person asking "what if I change this?"
That instinct is most of security testing. The OWASP Top 10 just tells you where to point it.
01What the OWASP Top 10 is
OWASP, the Open Worldwide Application Security Project, publishes a list of the ten most critical security risks for web applications, based on data from many organisations. The 2025 edition is the current one. Here it is, with what each item means in plain words and what a tester can check.
02The list, in plain words
| Risk | In plain words | What a tester can check |
|---|---|---|
| A01 Broken Access Control | People can see or do things they shouldn’t | Change ids in URLs and requests; open admin pages as a normal user; act on someone else’s data |
| A02 Security Misconfiguration | Settings left unsafe | Default passwords, detailed error pages, directory listings, debug features left on |
| A03 Software Supply Chain Failures | Problems in the libraries and tools you build with | Ask whether dependencies are scanned and updated; check the build pulls only trusted packages |
| A04 Cryptographic Failures | Sensitive data not properly protected | Pages served without HTTPS; passwords or card data visible in logs, URLs or exports |
| A05 Injection | Input treated as commands | Quotes, angle brackets and script-like text in every field; check they’re shown as text, not run |
| A06 Insecure Design | The feature itself is unsafe, however well coded | Unlimited coupon reuse, password reset by guessable questions, no limit on attempts |
| A07 Authentication Failures | Login and sessions can be abused | Lockout after failed attempts; logout really ends the session; reset links expire |
| A08 Software or Data Integrity Failures | Code or data trusted without checking | Can a user change a price or a role in data the app sends back to the server? |
| A09 Security Logging and Alerting Failures | Attacks happen and nobody notices | Are failed logins and permission errors logged? Does anyone get alerted? |
| A10 Mishandling of Exceptional Conditions | Errors handled in unsafe ways | Force errors (bad input, timeouts) and check the app fails safely without leaking details or skipping checks |
03Access control: the one to test every day
Broken access control is first on the list, and it is also the easiest for testers to find, because it needs no special tools. For every feature, ask: who should be allowed to do this, and have I tried it as someone who shouldn’t?
- Change identifiers. Order numbers, user ids and document ids in URLs and API requests. Try the next and previous number.
- Use the wrong role. Log in as a basic user and open admin URLs directly.
- Log out, then press Back. Can you still see private pages? Can you still perform actions?
- Send fields you shouldn’t. Add
"role":"admin"or"price":0to a request the app normally sends.
04Injection: test the input fields you already test
You already test fields with long text and odd characters. Add a few strings that would cause damage if the app treated them as code rather than text. The goal is not to break in; it is to check the app shows the input back as plain text and doesn’t behave differently.
| Try typing | A safe app | A warning sign |
|---|---|---|
| O’Brien (an apostrophe) | Saves and shows the name | An error page or a database message |
<b>bold</b> | Shows the tags as text | The text turns bold |
| A very long string | Rejects it with a message | A crash, or a server error |
If you see a warning sign, stop and report it. Proper injection testing on live or shared systems is for security specialists with permission.
05Know where your role ends
Testers catch many of these issues early, during normal work. That doesn’t replace a security review or penetration test. Only test systems you’re authorised to test, never use real customer data to probe for leaks, and report suspected vulnerabilities privately to the right people rather than in a public channel.
Security checks to add to everyday testing
- Every feature tested as a user who shouldn’t have access
- Ids in URLs and requests changed to another user’s
- Admin pages opened directly as a normal user
- Logout, Back and expired sessions checked
- Input fields tried with apostrophes, tags and very long text
- Error pages checked for stack traces or internal details
- Login lockout and reset link expiry tested
- Suspected vulnerabilities reported privately
Frequently asked questions
What is the OWASP Top 10?
A regularly updated list, published by OWASP, of the ten most critical security risks for web applications. The current edition is 2025.
What is number one in the OWASP Top 10 2025?
A01: Broken Access Control, meaning users can see or do things they shouldn’t be able to.
Can manual testers do security testing?
Yes, for many checks: access control, input handling, session behaviour and error messages. Deeper testing, such as penetration testing, needs specialists and explicit permission.
What changed in the 2025 OWASP Top 10?
Software Supply Chain Failures widens the old Vulnerable and Outdated Components category, Mishandling of Exceptional Conditions is new, and several categories moved. Broken Access Control stays at number one.
Is it legal to test for security problems?
Only on systems you own or are explicitly authorised to test. Testing someone else’s system without permission can be illegal, even with good intentions.
Written by
Vihari Panyam
QualityAIQ