Glossary · Agentic AI
What is Prompt-injection defense? Definition and meaning
Techniques that stop text in web pages, files or tool results from hijacking the agent.
Definition
Prompt-injection defense: Techniques that stop text in web pages, files or tool results from hijacking the agent.
In short: Protecting agents from hostile instructions in data.
When it is used
Safe use of agents on untrusted content.
Where it fits
Prompt-injection defense belongs to Operations in the Agentic AI topic. These are the other terms in the same group:
- Tracing: Captures prompts, tool calls, inputs and outputs of every step as a trace.
- Evals: Test sets and graders that measure quality of a prompt, model or agent.
- Regression boards: A standing set of benchmark rows that is re-run in full after each change so that a new fix cannot silently break an old result.
- Cost and latency budgets: Explicit ceilings on tokens, tool calls and seconds for an agent run.
- Sandboxing: Executes code and tools inside a restricted container or virtual machine.
Learn more
The full guide to Prompt-injection defense explains how it works, when to use it and how it compares with the alternatives, with examples. The glossary lists every term on the site from A to Z.
Previous term: Prompt tuning · Next term: PromptFoo