← Concept Index

Agent & LLM security

Prompt injection

Also called: instruction injection

DEFINITION

Prompt injection is when text the model reads — a web page, a document, an email — contains instructions that hijack what it does, overriding the task it was actually given.

WHY IT MATTERS

It's widely treated as the top security risk for LLM applications (it heads the OWASP Top 10 for LLM Applications). For an agent that can act, an injected 'forward these files' can become a real action, which is why consequential steps need a human gate.

COMMONLY CONFUSED WITH

A bug in the model. It's a structural problem: models can't reliably tell 'content to process' from 'instructions to follow', because both arrive as text.

SOURCES

First-pass citations, limited to primary sources; a reviewer will broaden and verify these before this entry leaves draft.