August 2026
-
Prompt Injection: The Vulnerability That Defines LLM Security
A model cannot tell its instructions from the data it reads. The direct form is a nuisance; the indirect form, hidden in content it ingests, is the real…
-
What AI Changes About Penetration Testing
AI changes penetration testing in two ways: a fast assistant that must always be checked, and a new class of target most organisations deploy faster than they secure.