ForesightSafety-SAGE:A Fully Automated Scenario Generation and Safety Evaluation Framework for LLM Agents

arXiv:2606.08531 · cs.AI · Submitted 2026-06-07 · Read on arXiv

cs.AI

Submitted: 2026-06-07

Updated: 2026-08-31

Comments: Accepted to EMNLP 2026 Main Conference

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Trace — 12× in this paper · explained in 2 episodes
• Benchmark — 9× in this paper · explained in 2 episodes
• Robustness — 6× in this paper · explained in 5 episodes
• Consistency — 5× in this paper · explained in 3 episodes
• LLM judge — 5× in this paper · explained in 2 episodes
• Attack Success Rate — 4× in this paper
• Adam — 3× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 3× in this paper · explained in 6 episodes
• Attack Success Rate (ASR) — 2× in this paper · explained in 2 episodes
• Generalization — 2× in this paper · explained in 13 episodes
• LLM-as-a-Judge — 2× in this paper · explained in 3 episodes
• Reward Hacking — 2× in this paper · explained in 8 episodes
• Adaptive Sampling — 1× in this paper · explained in 1 episode
• Agent Memory — 1× in this paper · explained in 1 episode
• Benchmarking — 1× in this paper · explained in 3 episodes
• Core — 1× in this paper · explained in 2 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Task Planning — 1× in this paper · explained in 2 episodes
• Unsafe Tool-call Rate — 1× in this paper
• Verification — 1× in this paper · explained in 2 episodes

Sources

• Mind the GAP: Text Safety Does Not Transfer to Tool-Call Safety in LLM Agents
• Towards Verifiably Safe Tool Use for LLM Agents
• GPT-4o System Card
• Kimi K2: Open Agentic Intelligence
• DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
• HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
• WebGPT: Browser-assisted question-answering with human feedback
• SafeToolBench: Pioneering a Prospective Benchmark to Evaluating Tool Utilization Safety in LLMs
• WizardLM: Empowering large pre-trained language models to follow complex instructions
• $\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains
• ReAct: Synergizing Reasoning and Acting in Language Models
• Agent-SafetyBench: Evaluating the Safety of LLM Agents

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems
• VSAL: A Vision Solver with Adaptive Layouts for Graph Property Detection