Resources

Despite headlines about autonomous LLM-driven cyberattacks, recent incidents like the ServiceNow Count(er) Strike vulnerability and so-called “LLM hijacking” campaigns all came down to old techniques: enumeration, poor ACLs, and exposed credentials.

We didn’t escalate privileges. We didn’t break anything.
We authenticated, then watched the CRM leak full names, departments, employee IDs, and account IDs into the browser.

Everything trusted the login.
And that trust is what got them compromised.

In this first installment of a real-world Lares engagement, we show how weak passwords, reused credentials, and login portal behavior enabled valid access to QA, sales, finance, and even the company’s founder—without phishing or exploiting a single vulnerability. Using only public data and internal credential leaks, we chained small wins into full authentication. This blog reveals how predictable employee behavior can bypass security controls long before an exploit is ever needed.

Solely on vulnerability scanning creates a false sense of security. Learn the limits of automated tools versus comprehensive, adversary-focused security testing for true cyber resilience.