Skip to content
ai

Prompt injection

Prompt Injection Attack

Definition

Prompt injection is a security vulnerability where malicious content in user inputs or retrieved documents overrides a system prompt's instructions, causing the AI to behave in unintended ways. Indirect prompt injection occurs when the malicious instruction is embedded in external data (web pages, documents) retrieved by an agent.

Defending against prompt injection requires input sanitization, privilege separation, and output monitoring.

Can this ship? Are we exposed? Can you prove it?

Every CISO is asking these three questions. We answer them with facts, not guesses, from repository to runtime.