ML Security Papers

LLM08

Excessive Agency

LLM agents acting beyond intended scope

219 papers Browse all papers

Monthly publications

Paper types

defense 89

benchmark 60

attack 37

survey 23

tool 10

Domains

nlp 216

multimodal 41

reinforcement-learning 12

vision 10

audio 3

graph 3

generative 1

Co-occurring categories

Other OWASP categories that appear on the same papers

LLM01 Prompt Injection

LLM07 Insecure Plugin Design

ML09 Output Integrity Attack

LLM06 Sensitive Information Disclosure

ML01 Input Manipulation Attack

ML10 Model Poisoning

ML06 AI Supply Chain Attacks

ML02 Data Poisoning Attack

ML07 Transfer Learning Attack

ML08 Model Skewing

Top cited papers

Agentic Misalignment: How LLMs Could Be Insider Threats

A-MemGuard: A Proactive Defense Framework for LLM-Based Agent Memory

Securing the Model Context Protocol (MCP): Risks, Controls, and Governance

Systematization of Knowledge: Security and Safety in the Model Context Protocol Ecosystem

A Survey on Agentic Security: Applications, Threats and Defenses

Agentic AI Security: Threats, Defenses, Evaluation, and Open Challenges

Securing AI Agent Execution

BrowseSafe: Understanding and Preventing Prompt Injection Within AI Browser Agents

Mitigating the OWASP Top 10 For Large Language Models Applications using Intelligent Agents

Spilling the Beans: Teaching LLMs to Self-Report Their Hidden Objectives

Browse all 219 papers