Class AdversarialResponseLibrary

java.lang.Object
org.mockserver.llm.adversarial.AdversarialResponseLibrary

public final class AdversarialResponseLibrary extends Object
A curated catalog of adversarial responses — the kind of hostile or malformed output an agent might receive from a compromised tool or a jailbroken upstream model. Mocking these as the LLM response lets you test that your agent resists them (does not leak its prompt, does not obey injected instructions, handles malformed/oversized content gracefully).

This is a defensive testing aid: the payloads are short, well-known example strings used as test fixtures, not working exploits. Purely deterministic generation — each id always maps to the same text.