Agent Skill Security Auditing

Overview

Agent skill security auditing studies whether reusable instructions, code, and tools behave as intended and stay within an agent’s authorized scope. Auditors combine static checks such as pattern matching and data-flow analysis with semantic review of natural-language instructions, sandboxed execution, and runtime monitoring. Benchmarks increasingly test complete agent trajectories, skill dependencies, and compositions, since individually benign skills can interact to cause unsafe actions. A central challenge is balancing high detection recall with false positives and reliable performance on unfamiliar skills, adaptive attacks, and different agent frameworks.

Latest papers

CardsList

72 papers9 papers in the last four weeks, down 25% on the four weeks before. 0.1% of all new papers.

  1. When Feedback Becomes a Backdoor: Tracing Backdoor Skill Formation in Text-Space Skill Optimization

    Nov 6, 2026Louise Nayeong Koh, Yaechan Lee, Myeongjong Lee +6Agent Skill Security AuditingAgent Skill Learning

  2. LatentTriad: Pre-Action Risk Representations for Malicious Agent Skill Detection

    Nov 3, 2026Zhe Yu, Yi Liu, Ying Zhang +1LLM Agent SecurityAgent Skill Security Auditing

  3. RuntimeTickets: Source-Grounded Security Obligations for Agent Skills

    Nov 2, 2026Kuiyi Gao, Xue Liu, Bowei HeLLM Agent SecurityRuntime Enforcement for AI Agents

  4. Scan Before You Install: Detecting and Mitigating Local and Composite Vulnerabilities in Skills

    Oct 27, 2026Parth A. Gandhi, Avital Aviv, Prasanna N. Wudali +2Agent Skill Security AuditingSoftware Vulnerability Detection

  5. Poisoned Skills: Propagating Jailbreaks Through Agent Markdown Files

    Oct 27, 2026Kashyap R Manjusha, Murali EmaniPrompt Injection Attacks on AI AgentsLLM Jailbreak Attacks

  6. When Skills Tell Agents Not to Ask: A Large-Scale Study of Oversight Directives in Agent Skills

    Oct 27, 2026Timofey Sanko, Yuan TianAgent Skill Security AuditingLLM Agent Safety

  7. SkillLifeBench: Benchmarking Security Across the Lifecycle of Agent Skills

    Oct 26, 2026Chenxu Zhang, Zhou Liu, Wu Xin +1AI Agent Security BenchmarksLLM Agent Evaluation

  8. CoSA: The Conflict of Skills Attack in Real-World LLM Agent Ecosystems

    Oct 26, 2026Yujia Zheng, Tianhao Li, Shangzhi Xu +6LLM Agent SecurityLLM Agent Safety

  9. SEEK: Skill Extraction with Evolving Knowledge against Frontier Science Agents

    Oct 25, 2026Cheng Gui, Xiang Zheng, Yutao Wu +5AI Agent SecurityAgent Skill Security Auditing

  10. CoRedSkill: Stealthy Skill Composition Attacks via AST-Guided Injection

    Oct 14, 2026Qichen Zhao, Shengfang Zhai, Qingni Shen +2AI Agent SecurityAutomated Red Teaming

  11. Beyond Maliciousness: Static and Runtime Attack-Surface Profiling of Agent Skills

    Oct 13, 2026Ratnesh Pandey, Japneet Singh, Harsh Akshit +2AI Agent SecurityAgent Skill Security Auditing

  12. Encrypted Skills Can Become New Security Exploits If AI Agents Know How to Decipher

    Oct 13, 2026Yan-Lun Chen, Pin-Yu Chen, Chia-Mu YuPrompt Injection Attacks on AI AgentsPrompt Injection Attacks

  13. SkillScreen: High-Recall Screening of Malicious Agent Skills with Decision Models

    Oct 13, 2026Jerrold Huang, Junwei Quan, Saiyue Lyu +2Agent Skill Security AuditingAI Agent Security Benchmarks

  14. Runaway Reaction: When Benign Skills Compose into Malicious Behavior

    Oct 5, 2026Zunlong Zhou, Ziyuan Yang, Mengyu Sun +1AI Agent SecurityLLM Agent Security

  15. Pretext: Defeating Malicious Skill Detection Frameworks for AI Agents

    Sep 30, 2026Tobias Kaiser, Aritra DharAI Agent SecurityAgent Skill Security Auditing

  16. Hiding in Plain Sight: Decoupling Pretext from Actuation for Skill Poisoning in LLM Agents

    Sep 30, 2026Wenxin Wu, Lingyong Yan, Lei Sha +2Adversarial Attacks on LLMsLLM Agent Security

  17. Can Agents Trust Their Skills? Uncovering Unsafe Chains of Trust in Skill-Based LLM Agents

    Sep 30, 2026Yan Wang, Zhihao Zhang, Ke Chen +5LLM Agent SecurityPrompt Injection Attacks on AI Agents

  18. SKILLLITE: Evidence-Guided Malicious Skill Auditing with Compact LLMs

    Sep 29, 2026Haoran Ou, Gelei Deng, Xuanye Zhang +3Small Language ModelsLLM Agent Security

  19. MMSkillRisk: Can Agents Stay Safe When Multimodal Skills Become Traps?

    Sep 28, 2026Lingqi Jiang, Jialuo Chen, Jianan Ma +8Prompt Injection Attacks on AI AgentsAgent Skill Security Auditing

  20. SkillDRE: Dual-Stage Red-Team Evolution of Agent Skills via Pre-Execution and Runtime Feedback

    Sep 26, 2026Pengyu Zhu, Jingyi Yang, Yi Liu +2Automated Red TeamingLLM Red Teaming

  21. Stealth Apart, Harm Together: Skill Cascading Attacks on Skill-Based Agent Systems

    Sep 24, 2026Zihao Zhu, Siwei Lyu, Adel Bibi +1Adversarial AttacksLLM Agent Security

  22. SkillAtlas: An Attack Trace Library for Agent Skills

    Sep 11, 2026Yuxin Tian, Zenghao Duan, Liang Pang +2LLM Agent SecurityAgent Skill Security Auditing

  23. AgentLeak: Cloning Stronger LLM Agent Capabilities onto Weaker Agents Beyond Skill Stealing

    Sep 7, 2026Xiaoting Lyu, Yuhong Wu, Yufei Han +6LLM Agent SecurityLLM Agent Skill Learning

  24. Safe Skill Retirement for Physical Agents

    Aug 25, 2026Zhonghao Zhan, Xiao Ma, Hamed HaddadiLLM Agent SafetyAgent Skill Security Auditing

  25. ClawSentry: A Progressive Multi-Tier Security Monitor for Safeguarding Autonomous LLM Agents

    Aug 21, 2026Kai Wang, Zeming Wei, BiaoJie Zeng +7AI Agent SecurityAI Agent Monitoring

  26. MaliciousSkillBench: A Comprehensive Benchmark for Malicious Agent Skill Detection

    Aug 20, 2026Yue Wang, Yi Liu, Gelei Deng +4AI Agent Security BenchmarksAgent Skill Security Auditing

  27. TRUSS: Towards Task-Reliable and User-Safe Automated Agent Skill Generation

    Aug 18, 2026Zhibo Zhang, Zhen Ouyang, Ling Shi +1Agent Skill Security AuditingTool Access Control for LLM Agents