AI Agent Security: 4 Failure Modes That Are Not Prompt Injection

Description and use

<p>Prompt injection is one item on a list of ten.</p> <p>That is worth sitting with, because if you read the agent security conversation online you would think it was the whole list. The <a href="https://genai.owasp.org/llm-top-10/" rel="noopener noreferrer">OWASP Top 10 for LLM Applications 2025</a> opens with LLM01:2025 Prompt Injection, and then names nine more things that will take your system down: Sensitive Information Disclosure, Supply Chain, Data and Model Poisoning, Improper Output Handling, Excessive Agency, System Prompt Leakage, Vector and Embedding Weaknesses, Misinformation, and

Problem

<p>Prompt injection is one item on a list of ten.</p> <p>That is worth sitting with, because if you read the agent security conversation online you would think it was the whole list. The <a href="https://genai.owasp.org/llm-top-10/" rel="noopener noreferrer">OWASP Top 10 for LLM Applications 2025</a> opens with LLM01:2025 Prompt Injection, and then names nine more things that will take your system down: Sensitive Information Disclosure, Supply Chain, Data and Model Poisoning, Improper Output Handling, Excessive Agency, System Prompt Leakage, Vector and Embedding Weaknesses, Misinformation, and

Outcome

<p>Prompt injection is one item on a list of ten.</p> <p>That is worth sitting with, because if you read the agent security conversation online you would think it was the whole list. The <a href="https://genai.owasp.org/llm-top-10/" rel="noopener noreferrer">OWASP Top 10 for LLM Applications 2025</a> opens with LLM01:2025 Prompt Injection, and then names nine more things that will take your system down: Sensitive Information Disclosure, Supply Chain, Data and Model Poisoning, Improper Output Handling, Excessive Agency, System Prompt Leakage, Vector and Embedding Weaknesses, Misinformation, and

Evidence

AI Agent Security: 4 Failure Modes That Are Not Prompt Injection

Source