Concept  ·  Glossary

Prompt injection taxonomy

A structured, named catalog of the many different ways attackers can smuggle hidden instructions into an AI system to hijack its behavior — now expanded by security vendors to over 200 distinct documented techniques. Having a shared taxonomy lets defenders name and track specific attack variants the way antivirus companies name malware families.
The sheer size of this catalog (200+ techniques and growing) signals that prompt injection isn't a single bug to patch but an entire, rapidly expanding attack category that security teams must continuously monitor.
Track this in the live feed See how this plays out in real AI security and governance developments.
Open the feed →