guardscope: Measuring Which Actions Runtime Controls Check in LLM Agent Systems
cs.AI
Submitted: 2026-11-11
Updated: 2026-11-11
Code: https://github.com/VulcanLab/guardscope
Project page: https://microsoft.github.io/agentgovernance-toolkit
Terminology
Sources
- Design Patterns for Securing LLM Agents against Prompt Injections
- ShieldAgent: Shielding Agents via Verifiable Safety Policy Reasoning
- LlamaFirewall: An open source guardrail system for building secure AI agents
- Causality Laundering: Denial-Feedback Leakage in Tool-Calling LLM Agents
- Cedar: A New Language for Expressive, Fast, Safe, and Analyzable Authorization (Extended Version)
- Defeating Prompt Injections by Design
- AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and Defenses for LLM Agents
- AgentOps: Enabling Observability of LLM Agents
- Building Guardrails for Large Language Models
- Safeguarding Large Language Models: A Survey
- Securing AI Agents: Implementing Role-Based Access Control for Industrial Applications
- Distributing Security Controls Through Harness Engineering
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- Taming Various Privilege Escalation in LLM-Based Agent Systems: A Mandatory Access Control Framework
- APPA: Recoverable Information-Flow Control for Real-World LLM Agents
- Instruction Self-Arbitration Is Not a Trust Boundary: Configuration-Dependent Tool Calls and External Enforcement in LLM Agents
- Visual Confused Deputy: Exploiting and Defending Perception Failures in Computer-Using Agents
- AgentGuard: An Attribute-Based Access Control Framework for Tool-Use LLM-Based Agent
- Adaptive Evaluation of Out-of-Band Defenses Against Prompt Injection in LLM Agents
- Formal Policy Enforcement for Agentic Systems
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- 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
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents