VEPA: Verified Evolving Formal Policy for Lifelong LLM Agent Guardrails
cs.AI, cs.CL, cs.CR, cs.LG
Submitted: 2026-10-08
Updated: 2026-10-08
Comments: 33 pages, 14 figures
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- AgentGuardian: Learning Access Control Policies to Govern AI Agent Behavior
- LlamaFirewall: An open source guardrail system for building secure AI agents
- Membrane: A Self-Evolving Contrastive Safety Memory for LLM Agent Defense
- Agent Security is a Systems Problem
- Securing AI Agents with Information-Flow Control
- Don't Make Models Guess Security and Safety: Symbolic Guardrails for Domain-Specific AI Agents
- Memory in the Age of AI Agents
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- The Attack and Defense Landscape of Agentic AI: A Comprehensive Survey
- LiSA: Lifelong Safety Adaptation via Conservative Policy Induction
- Inductive Learning of Answer Set Programs from Noisy Examples
- Formal-LLM: Integrating Formal Language and Natural Language for Controllable LLM-based Agents
- Large Language Model Agent: A Survey on Methodology, Applications and Challenges
- VeriGuard: Enhancing LLM Agent Safety via Verified Code Generation
- Illuminating search spaces by mapping elites
- AlphaEvolve: A coding agent for scientific and algorithmic discovery
- Formal Policy Enforcement for Agentic Systems
- Progent: Securing AI Agents with Privilege Control
- AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents
- SkillOpt: Executive Strategy for Self-Evolving Agent Skills
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
- Agents' Last Exam
- 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