论文 AID-Guard: Stateful Authorization for Delegated Agent Effects
论文 Beyond Success Rate: Cost-Aware Evaluation of Offensive and Defensive Security Agents
论文 4️⃣ arXiv · Securing the Agent: Vendor-Neutral, Multitenant Enterprise Retrieval and Tool Use(⭐⭐⭐⭐ 高优先级)
论文 5. Context-Fractured Decomposition Attacks on Tool-Using LLM Agents
论文 From Detection to Action: Using LLM Agents for Fault-Tolerant Control
论文 Behavioral Privacy Leakage in Agentic Negotiation: Formalizing and Mitigating Inference Attacks via Randomized Policies