Indirect prompt injection
Indirect prompt injection is prompt injection delivered through content the model retrieves itself — a web page, an email, a document — rather than typed by the user.
It is the more serious variant because no one involved intends to attack anything: a user asks an agent to summarise a page, and the page contains hidden instructions. Any agent that both reads untrusted content and can take actions needs a boundary between the two.
Related terms
- Prompt injection
- Prompt injection is an attack in which instructions hidden inside content a model processes cause it to ignore its original instructions.
- AI agent
- An AI agent is a language model given tools and a goal, which decides for itself which actions to take and in what order until the goal is met.
More on safety
See the full glossary, read the guides, or put it into practice in the prompt builder.