Security testing basics
The most common web vulnerabilities, how a QA engineer can check for them safely, and where professional penetration testing begins.
Security is a quality characteristic like any other, and many serious vulnerabilities are found by ordinary testers who simply ask “what if I'm not supposed to be able to do this?”. You don't need to be a hacker to catch the most common problems — you need curiosity, DevTools and a clear understanding of what the system should forbid.
The most common weaknesses
The OWASP Top 10 is a widely used list of the most critical web application risks. In everyday QA work, these are the ones you meet most often:
| Risk | What goes wrong | A simple tester's check |
|---|---|---|
| Broken access control | Users reach data or actions that aren't theirs | Change an ID in the URL or API request; open admin URLs as a regular user |
| Injection | User input is executed as code or a query | Enter quotes, <script> tags and SQL fragments in fields; check they are shown as text |
| Authentication failures | Weak login, sessions that never expire | Brute-force limits, logout really ends the session, password reset links expire |
| Sensitive data exposure | Secrets or personal data leak | Look at API responses in DevTools for extra fields: password hashes, emails of others, tokens |
| Security misconfiguration | Debug pages, default passwords, verbose errors | Trigger errors and check no stack traces or server versions appear |
| Vulnerable components | Old libraries with known holes | Ask whether dependency scanning runs in CI |
Access control: the tester's favourite
Most applications check permissions in the interface but forget to check them on the server. Hiding a button is not security. A practical routine:
- Prepare two users with different roles (for example, a customer and an admin) and two customers with their own data.
- As user A, perform an action and copy the request from DevTools (right-click → Copy as cURL).
- Replay it with user B's session, or change the object ID to one that belongs to user A.
- Expected result:
403or404and no data. Anything else is a critical bug.
Input handling
Every field that accepts text is a potential entry point. Try values like these and check that the application stores and shows them literally, without breaking the page or the query:
<script>alert(1)</script>
"><img src=x onerror=alert(1)>
' OR '1'='1
Robert'); DROP TABLE students;--
../../etc/passwd
{{7*7}}
If an alert pops up, the page layout breaks, or {{7*7}} turns into 49, report it immediately. These are signs that input reaches places where it is interpreted as code.
Sessions, cookies and headers
- Session cookies should have
HttpOnly,Secureand a sensibleSameSitevalue - After logout, pressing Back or replaying an old request must not show private data
- Changing a password should end other active sessions
- Responses should include basic security headers such as
Content-Security-Policy,Strict-Transport-SecurityandX-Content-Type-Options - Error pages must not reveal stack traces, SQL queries or server versions
Tools
| Tool | Use |
|---|---|
| Browser DevTools | Inspect requests, cookies, storage, headers |
| OWASP ZAP | Free proxy and scanner for web applications |
| Burp Suite (Community) | Intercept and modify requests |
| Dependency scanners (npm audit, Dependabot, Snyk) | Known vulnerabilities in libraries |
| Security headers checkers | Quick review of HTTP response headers |
Where QA ends and pentesting begins
Testers cover the basics continuously: access control, input handling, sessions, configuration. Systems that handle payments, health or other sensitive data also need regular professional penetration tests and code audits. The two complement each other — everyday checks catch regressions between audits.


