Prompt injection

Malicious instructions hidden in content the agent reads (web page, email, doc) that trick it into harmful actions. The defining agent security threat.

Malicious instructions hidden inside content the model reads (a document, a web page, a tool result) that try to hijack its behavior. The defining LLM security threat.

Hidden malicious instructions inside a retrieved document that try to hijack the model; why retrieved content is untrusted.

An attack where malicious instructions in the text the model reads hijack its behavior. The signature AI attack. (Mod 6)

Malicious instructions hidden in input (especially dangerous with tools or untrusted RAG documents).

An attack where malicious instructions hidden in content the agent reads hijack its behavior; includes the "Rules File Backdoor" and RoguePilot-class attacks. (Mod 15)