When “Working as Designed” Becomes the Problem
Take Stripe.
In a publicly discussed incident, attackers exploited weaknesses in API integrations and business logic to carry out fraudulent transactions. The APIs themselves were functional and authenticated—the issue was how workflows could be abused in sequence, at scale.
Or consider McDonald's.
Researchers discovered that a hiring chatbot API exposed sensitive applicant data due to weak access controls and predictable request patterns. By interacting with the API in unintended ways, they were able to retrieve personal information from job applicants.
These weren’t traditional “break-ins.”
They were cases of APIs behaving correctly—but being used in ways no one anticipated.
That’s exactly the category of risk most organizations struggle to see.
Why Visibility Is Harder Than It Sounds
Ask most security or DevSecOps teams what their API inventory looks like, and you’ll get an incomplete answer.
Not because they’re careless—but because the environment is constantly changing.
Effective API discovery isn’t a one-time scan or a manually maintained list. It has to be:
- Continuous: capturing APIs as they appear, change, and disappear
- Automated: without relying on developers to keep documentation updated
- Comprehensive: covering internal, external, third-party, shadow, and deprecated APIs
But just finding endpoints isn’t enough.
To actually understand risk, you need to go deeper:
- What parameters does each API accept?
- What does the request and response body look like in practice?
- How does the API behave across different inputs and conditions?
This is where advanced APM solutions stand out—they don’t just enumerate endpoints, they reconstruct accurate API schemas based on real traffic.
And even that is only part of the picture.