Website, Web Application and API Penetration Testing
Your website, customer portal and APIs are the part of the organization anyone on the internet can reach, and often also the part that holds customers’ personal data. A web application penetration test checks them against OWASP: whether one customer can reach another customer’s data, whether authentication holds up, and whether admin interfaces, files or keys were left exposed.
A system open
to the entire internet
We test the system from the outside, as an anonymous visitor (unauthenticated), and from the inside, with test users at different permission levels (authenticated). That’s how weaknesses an automated scanner can’t see come to light: permissions between users, business logic and processes that can be bypassed.
- OWASP Top 10, OWASP API Security Top 10 and OWASP WSTG
- Testing with and without users (authenticated / unauthenticated)
- Preference for a staging environment, and full caution in production
Who it’s for
Businesses with a customer portal or personal area
Where customers, patients or suppliers log in with a username and password and see personal data, documents or invoices.
Companies developing a system or app
With APIs for an app, partners or integrations, that want to test the system before go-live or after a major change.
Websites on WordPress or off-the-shelf systems
With plugins, forms and admin areas built over years by several providers, without an orderly security review.
Sound familiar?
These situations come up again and again in websites and systems built under deadline pressure.
- It’s unclear whether one customer can reach another customer’s documents or orders (IDOR)
- An admin panel reachable from the entire internet, protected by a password alone
- A WordPress site with plugins that haven’t been updated, some of them no longer maintained
- API keys and passwords left in code that runs in the browser (front-end)
- A login form with no limit on the number of attempts (rate limiting)
- The system was built by a development vendor, and nobody checked its security before it went live
What a web and API test checks
The scope depends on the system, the types of users and the data it holds, with written authorization and agreed rules of engagement.
Authentication and session management
Login, password reset, MFA, session expiry, logout and resistance to password guessing (brute force).
Permissions and access control
What each type of user is allowed to see and do, and whether one user can reach another user’s data (IDOR) or admin actions.
Injection
Testing input points, such as forms, parameters and APIs, for injection into the database (SQL injection), into system commands and into the browser (XSS).
Business logic
Processes that can be bypassed or used other than as intended, for example skipping a step in a payment process or reusing a coupon or an approval.
File upload
Which file types are allowed, where they are stored, whether the server might execute them, and whether one user’s files are accessible to others.
APIs
Testing against the OWASP API Security Top 10: object-level authorization (BOLA), returning excess data, rate limiting and old versions left open.
WordPress, configuration and exposure
Outdated plugins and themes, exposed admin interfaces, backup and configuration files forgotten on the server, and security headers.
Report, remediation and retest
Findings rated by CVSS, a fix recommendation (remediation) that developers can implement, and a retest after the fix.
What we take care of, in detail
Application and API
- Methodology: OWASP WSTG, with findings mapped to the OWASP Top 10 and the OWASP API Security Top 10
- Authentication: login, registration, password reset, MFA, SSO and OAuth
- Session management: cookies, expiry, logout and CSRF protection
- Access control between users and between roles: IDOR / BOLA, privilege escalation and access to admin actions
- Injection: SQL, NoSQL, system commands, XSS and SSRF
- Business logic: payment, ordering, approval and workflow processes
- File upload: type, size, storage location and access permissions
- REST and GraphQL APIs: returning excess data, rate limiting, old versions and exposed documentation
- Secrets in client-side code (front-end): API keys, tokens and addresses of internal environments
Server, configuration and WordPress
- WordPress: core, plugins and themes, admin users, XML-RPC and REST API
- Exposed admin panels: the site’s admin interface, database management and server control panels
- Files forgotten on the server: backups, configuration files, version control folders and logs
- Security headers: CSP, HSTS, X-Frame-Options and secure cookies
- TLS settings and SSL certificates
- Error messages that reveal technical details (information disclosure)
- Web application firewall (WAF) and rate limiting: whether they exist, and what they actually block
How we work
- Written authorization from the system owner, and if it is hosted by a provider, checking the provider’s testing policy in advance
- Scope definition: addresses, environments, APIs and user roles, and agreed rules of engagement
- Testing without a user (unauthenticated) and with test users in every role (authenticated)
- Black, grey or white box, including an option for a focused code review
- Preference for a staging environment identical to production, and in production only non-destructive methods
- A coordinated time window, an available contact person, and a working backup of the site and the database before the test starts
- Working with test data rather than real customer data, wherever possible
From the first conversation to an organized routine
- 1
Introductory call
What the system does, who the users are, what data it holds and where it is hosted.
- 2
Scope, authorization and rules of engagement
Addresses, environments, APIs and user roles, written authorization and agreed rules of engagement.
- 3
Running the test
Testing with and without test users, preference for a staging environment, and updating the contact person during the test about any critical finding.
- 4
Report, remediation and retest
Findings rated by CVSS, remediation on the server and in the code, and a retest.
What a human tester finds that an automated scanner misses
Automated scanners are good at finding old versions and missing settings. Many of the serious weaknesses in websites and systems have to do with permissions and logic, and those are tested with an understanding of the system:
Permissions between users (IDOR)
Broken access control is at the top of the OWASP Top 10 list. The question is whether the server checks, on every request, that the requested data belongs to the logged-in user, and this is tested with several test users at different permission levels, because a scanner doesn’t know whose invoice is whose. Personal data exposed this way is a security incident (data breach) with consequences under the Privacy Protection Law and Amendment 13.
Business logic
A business logic flaw is a process that works exactly as it was written, but can be used in a way nobody planned: skipping a step, repeating an action or changing a value that was supposed to be fixed. It has no signature, so it is found only when you understand what the system is supposed to do. This is one of the reasons a test with test users (authenticated) finds more than a test from the outside.
APIs are tested directly
Mobile apps, SPA systems and partner integrations work against an API, and sometimes the API returns more data than the screen shows. The OWASP API Security Top 10 focuses on problems that recur in such interfaces, such as object-level authorization (BOLA), weak authentication and missing rate limiting. That’s why the API is tested directly, including with requests the website itself never sends.
WordPress and server configuration
On WordPress sites the weaknesses are mostly in plugins and themes, and less in the core. A plugin whose developer stopped maintaining it, an admin interface open to the entire internet or a backup file forgotten on the server can expose the site and the database. Some of the fixes are made on the server and hosting side: security headers, restricting access to the admin interface and a web application firewall (WAF).
Staging, production and who fixes what
It’s better to run a web test in a staging environment identical to production, because there you can test forms, payments and uploads without affecting real customers. When testing in production, we use non-destructive methods, in a coordinated time window and after a working backup. When we also manage the hosting and the server, we make the configuration fixes, and code findings go to the developers with a clear recommendation and a retest.
Guide: Penetration testing for websites and web systems: what OWASP says to test
Frequently asked questions
About penetration testing for websites, web applications and APIs.
Didn’t find an answer? +972-3-3801180
A security plugin and a web application firewall (WAF) block some of the common attacks, and they are an important layer. They don’t know whether one customer can reach another customer’s data (IDOR), whether the payment process can be bypassed, or whether an API returns more data than it should. Those are checked in a test that understands the system.
Testing without a user (unauthenticated) shows what anyone who reaches the site from the internet sees. Testing with test users (authenticated) in every role, for example customer, representative and admin, shows what can be done after logging in, and that is usually where the serious weaknesses are. In most cases it’s worth testing both.
Preferably in a staging environment identical to production, with test data. That way forms, payments and file uploads can be tested without affecting real customers and real data. When there is no staging environment, we test in production with non-destructive methods, in a coordinated time window, with an available contact person and after a working backup of the site and the database.
Yes, with written authorization from the site owner and in line with the hosting or cloud provider’s testing policy, which we check in advance. If the site is hosted with us, coordination is simpler and we make the configuration fixes on the server ourselves. Third-party SaaS systems that the organization only uses are not included in the scope without their provider’s consent.
We fix configuration, server, hosting and WordPress findings when they are in infrastructure we manage. Findings in the system’s code go to the developers, yours or the vendor’s, with a clear description of the problem and of the recommended fix (remediation). After the fix we carry out a retest to make sure the finding is closed.
Before a new system goes live, after a significant change to authentication, permissions or the payment process, and periodically on websites and systems that hold personal data. A requirement from a customer, an insurer or a regulator is also a reason to test, and where personal data is involved, the Privacy Protection Law and the Data Security Regulations apply.
More you can centralize with us
Information security and cyber for organizations
An outsourced CISO, risk management and compliance, hardening, penetration tests, monitoring and response to cyber incidents (incident response).
Penetration Testing
A controlled test of the network, websites, cloud and employees from an attacker’s point of view, with a findings report rated by risk (CVSS), fix recommendations (remediation) and a retest.
Website hosting and web servers
Website hosting, managed WordPress and WooCommerce on cPanel servers, with security, performance, SSL, backup and website migration between providers.
Want an in-depth test of your website or system?
Tell us briefly about your business and what concerns you about your IT. We’ll get back to you to understand what you need and suggest a way of working that fits.
Business hours: Sunday to Thursday, 08:00 to 17:00
Existing client? Open a support ticket