Trust Is Not a Score: Runtime Assurance Contracts for High-Risk AI Agents
cs.AI, cs.CY, cs.SE
Submitted: 2026-09-30
Updated: 2026-09-30
Comments: 16 pages, 2 figures, 5 tables. Ancillary files: decision log, executable transition model, LLM-labelled synthetic holdout. Synthetic mechanism study; no deployment claim
Code: https://github.com/SZabolotnii/TRACE-RAC-code-supplement
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Assurance Cases as Foundation Stone for Auditing AI-enabled and Autonomous Systems: Workshop Results and Political Recommendations for Action from the ExamAI Project
- TraceSafe: A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories
- Towards Verifiably Safe Tool Use for LLM Agents
- Selective Classification for Deep Neural Networks
- Guidance on the Assurance of Machine Learning in Autonomous Systems (AMLAS)
- Safe Reinforcement Learning via Probabilistic Shields
- Tiered Agentic Oversight: A Hierarchical Multi-Agent System for Healthcare Safety
- Online Shielding for Stochastic Systems
- Correct-by-Construction Runtime Enforcement in AI -- A Survey
- ToolSafe: Enhancing Tool Invocation Safety of LLM-based agents via Proactive Step-level Guardrail and Feedback
- A Theory of Least Autonomy in AI
- Easy-to-Use Shielding for Reinforcement Learning
- Progent: Securing AI Agents with Privilege Control
- GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning
- Do Coding Agents Understand Least-Privilege Authorization?
- TRACE: A Metrologically-Grounded Engineering Framework for Trustworthy Agentic AI Systems in Operationally Critical Domains
- MiniScope: Authorizing Agents with Least-Privilege Permissions
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