Authorization Drift in LLM Agents: Evaluating Boundary Compliance in Multi-Step Tool Use
cs.CR, cs.LG
Submitted: 2026-11-06
Updated: 2026-11-06
Code: https://github.com/ZJR-FZD/AgentDrift
Terminology
Sources
- The Llama 3 Herd of Models
- When Benign Inputs Lead to Severe Harms: Eliciting Unsafe Unintended Behaviors of Computer-Use Agents
- EU-Agent-Bench: Measuring Illegal Behavior of LLM Agents Under EU Law
- Ministral 3
- Progent: Securing AI Agents with Privilege Control
- IsolateGPT: An Execution Isolation Architecture for LLM-Based Agentic Systems
- On the Tool Manipulation Capability of Open-source Large Language Models
- Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward
- $\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains
- ReAct: Synergizing Reasoning and Acting in Language Models
- Evaluating Privilege Usage of Agents with Real-World Tools
- Agent-SafetyBench: Evaluating the Safety of LLM Agents
Related papers
- From Strategy Search to Policy Learning: MCTS-Guided Adversarial Co-Evolution for LLM Safety Alignment
- Privacy Leakage from Recurrent State in LLM Inference
- Incrementally Verifiable Oracle Aided Computation
- Privilege and confidentiality in generative AI workflows
- BadMoE: Backdooring Mixture-of-Experts LLMs via Routing-Aware Triggers and Task-Underutilized Experts
- CoVault: Secure, Scalable Analytics of Personal Data