RePolicy: Reinforcement Learning for Safety-Policy Invocation in Agent Safeguards

arXiv:2608.24275 · cs.AI, cs.CL · Submitted 2026-08-25 · Read on arXiv

cs.AI, cs.CL

Submitted: 2026-08-25

Updated: 2026-08-27

Code: https://github.com/jianghoucheng/RePolicy

Terminology

• Benchmark — 61× in this paper · explained in 2 episodes
• Reinforcement Learning — 43× in this paper · explained in 4 episodes
• GRPO — 21× in this paper · explained in 6 episodes
• Trace — 7× in this paper · explained in 2 episodes
• Fine-Tuning — 6× in this paper · explained in 5 episodes
• Ablation — 4× in this paper · explained in 1 episode
• Accuracy — 4× in this paper · explained in 1 episode
• Grounding — 4× in this paper · explained in 3 episodes
• AgentAuditor — 2× in this paper · explained in 1 episode
• Agentic AI — 2× in this paper · explained in 1 episode
• Aggregation — 2× in this paper · explained in 1 episode
• Robustness — 2× in this paper · explained in 5 episodes
• Verifiable Rewards — 2× in this paper · explained in 2 episodes
• Adam — 1× in this paper · explained in 2 episodes
• Adaptability — 1× in this paper · explained in 1 episode
• Adaptation — 1× in this paper · explained in 1 episode
• Adaptive Safety — 1× in this paper · explained in 1 episode
• Agentic Reasoning — 1× in this paper · explained in 1 episode
• Alignment — 1× in this paper · explained in 1 episode
• Benchmarking — 1× in this paper · explained in 3 episodes
• Calibration — 1× in this paper · explained in 7 episodes
• Consistency — 1× in this paper · explained in 3 episodes
• Error Decomposition — 1× in this paper · explained in 3 episodes
• GRPO Objective — 1× in this paper · explained in 2 episodes
• Generative AI — 1× in this paper · explained in 2 episodes
• Interpretability — 1× in this paper · explained in 15 episodes
• KL Regularization — 1× in this paper · explained in 2 episodes
• Large Language Model (LLM) — 1× in this paper · explained in 3 episodes
• Mixture-of-Experts (MoE) — 1× in this paper · explained in 7 episodes
• Mixture-of-Experts (MoE) Models — 1× in this paper · explained in 3 episodes
• Regularization — 1× in this paper · explained in 3 episodes
• Reward Function — 1× in this paper · explained in 2 episodes
• Self-consistency — 1× in this paper · explained in 2 episodes
• Transformer — 1× in this paper · explained in 2 episodes

Sources

• DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
• ReTool: Reinforcement Learning for Strategic Tool Use in LLMs
• AgentHazard: A Benchmark for Evaluating Harmful Behavior in Computer-Use Agents
• DynaGuard: A Dynamic Guardian Model With User-Defined Policies
• Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
• The Attack and Defense Landscape of Agentic AI: A Comprehensive Survey
• LPG: Balancing Efficiency and Policy Reasoning in Latent Policy Guardrails
• ATBench: A Diverse and Realistic Agent Trajectory Benchmark for Safety Evaluation and Diagnosis
• AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Security
• AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security
• AgentAuditor: Human-Level Safety and Security Evaluation for LLM Agents
• Large Language Model Agent: A Survey on Methodology, Applications and Challenges
• OpenAI GPT-5 System Card
• Towards trustworthy agentic AI: a comprehensive survey of safety, robustness, privacy, and system security
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
• Agents of Chaos
• OpenAgentSafety: A Comprehensive Framework for Evaluating Real-World AI Agent Safety
• HINTBench: Horizon-agent Intrinsic Non-attack Trajectory Benchmark
• A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment
• GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning

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