The model can be right and the deployment still wrong. Runtime settings, constants and environment configuration sit outside the app and can undo what it enforces, and the most common way in is not production at all: it is the acceptance environment that shares a secret with it.
Shipped defaults are public knowledge, and they are the first thing an attacker tries.
The same security-sensitive constant in test, acceptance and production, so the weakest environment sets the security of the strongest.
Real data in a non-production database, protected by whatever the test environment happens to have.
Missing security headers, or a content security policy loose enough that an XSS reaches as far as it likes.
API documentation and internal paths restricted on one environment and open on the next.
Multi-session, duration and timeout left on whatever the defaults were, regardless of what the application holds.
Every item below has been found in production applications. None of them requires a sophisticated attacker: they require somebody who reads the page source, edits a request or simply logs in.
Security-sensitive constants and environment variables that differ everywhere, so one environment cannot lend its keys to another.
No production credentials, APIs or data outside production, and anonymised data where test data is needed.
Path restrictions, headers and a strict content security policy applied per environment, and checked before each deploy.
Deployment configuration sits outside the model, so Omnext reports what it can see from the application side and the rest belongs in your release checklist. It is the one category where automation stops and process takes over.
The S-Unit Top 10 is the work of The S-Unit, the Mendix security specialists. They built it from more than ten years of offensive security work and hundreds of Mendix penetration tests, and they maintain the canonical version. Omnext automates the checks in the Security Analysis module; the framework itself is theirs, and this page explains it in our own words.
Ten questions, about four minutes, results on screen. Or let the scan answer all ten with evidence from your own model.