Similar Items: No More, No Less: Task Alignment in Terminal Agents
- LoopTrap: Termination Poisoning Attacks on LLM Agents
- When Alignment Isn't Enough: Response-Path Attacks on LLM Agents
- Safety Context Injection: Inference-Time Safety Alignment via Static Filtering and Agentic Analysis
- STARE: Step-wise Temporal Alignment and Red-teaming Engine for Multi-modal Toxicity Attack
- You Snooze, You Lose: Automatic Safety Alignment Restoration through Neural Weight Translation
- Stateful Agent Backdoor