What question did this study set out to answer?

The aim is to create a secure AI agent framework that addresses vulnerabilities in traditional probabilistic authorization systems.

March 12, 2026Open Access

PROVA: Proof-based Reasoning for Operating Verified Agents — A Formally Verified Pre-Execution Safety Gate for AI Agent Frameworks

Key Points

The aim is to create a secure AI agent framework that addresses vulnerabilities in traditional probabilistic authorization systems.
Developed the PROVA framework incorporating a formally verified pre-execution safety gate.
Replaced traditional probabilistic authorization with OR1ON's prove() primitive.
Designed to secure actions of AI agents against vulnerabilities like CVE-2026-25253.
Achieved 0 false positives on critical actions.
Demonstrated an average safety gate latency of 0.09ms.
Ensured 100% detection of prompt injection attacks.
Provided a full Merkle-signed audit trail for accountability.
Achieved compliance with the EU AI Act Hochrisiko by design.

Abstract

We present PROVA (Proof-based Reasoning for Operating Verified Agents), the first AI agent framework built around a formally verified pre-execution safety gate. In response to CVE-2026-25253 (CVSS 8.8, OpenClaw one-click RCE, 42,000+ affected instances), we demonstrate that the root cause of AI agent security vulnerabilities is architectural: probabilistic LLMs cannot reliably authorize agent actions. PROVA replaces probabilistic authorization with OR1ON's prove() primitive. Every agent action is formally proven safe before execution, or the system abstains. Key results: 0 false positives on critical actions, 0.09ms average safety gate latency, 100% prompt injection detection, full Merkle-signed audit trail, EU AI Act Hochrisiko compliant by design.

Read Full Paperexternally

AI에게 질문

Bookmark

View Full Paper