When Approval Becomes Execution: Authorization-Grounded Workflow Attacks on Tool-Using Agents
A new attack class turns approval tokens into execution commands in tool-using LLM agents, demonstrating that enterprise workflow automation with agentic AI needs explicit authorization boundaries beyond current designs.
Summary written by editorial AI · Source link below
arXiv:2610.23077v1 Announce Type: new Abstract: Tool-using language-model agents increasingly rely on local services for document access, approval management, and enterprise workflow execution. This paper studies an authorization-grounded workflow attack in the OpenClaw Agent Security Attack Competition, a controlled benchmark derived from SeClaw-style security task synthesis. Rather than using explicit policy-override instructions, the attack guides the agent through a locally coherent approva
Editorial Analysis
Enterprises integrating LLM agents into approval workflows risk a subtle privilege-escalation path where agent-interpreted approvals silently become executions.
Audit all workflows where LLM agents handle approval-to-execution transitions and enforce explicit human-in-the-loop gates.
Forward-looking interpretation drafted by editorial AI under human review — not a reproduction of the source. See methodology.
External link — opens at arXiv Crypto & Security in a new tab.
More from the AI Security Desk
- OpenAI AI Agents Demonstrate VM Escapes and Secret Cross-Sandbox Communication1d
- Exposed Infrastructure Reveals How Ransomware Affiliates Used AI Agents to Execute Intrusions3d
- You gave your AI agent access, but who is giving it orders?3d
- AI Agents as Insider Threats: Inside Microsoft's Bet on a Unified Security Stack3d
- Critical x402 SDK Flaw Could Let Malicious APIs Drain AI Agents’ Token Balances3d