AIJan 14

CaMeLs Can Use Computers Too: System-level Security for Computer Use Agents

arXiv:2601.09923v18 citationsHas Code
Originality Incremental advance
AI Analysis

This addresses security risks for AI agents automating tasks on computers, though it is incremental as it builds on known isolation principles.

The paper tackles the vulnerability of Computer Use Agents (CUAs) to prompt injection attacks by proposing Single-Shot Planning, which generates a complete execution graph before observing potentially malicious UI content to ensure security. It retains up to 57% of the performance of frontier models and improves smaller models by up to 19% on OSWorld, showing that security and utility can coexist.

AI agents are vulnerable to prompt injection attacks, where malicious content hijacks agent behavior to steal credentials or cause financial loss. The only known robust defense is architectural isolation that strictly separates trusted task planning from untrusted environment observations. However, applying this design to Computer Use Agents (CUAs) -- systems that automate tasks by viewing screens and executing actions -- presents a fundamental challenge: current agents require continuous observation of UI state to determine each action, conflicting with the isolation required for security. We resolve this tension by demonstrating that UI workflows, while dynamic, are structurally predictable. We introduce Single-Shot Planning for CUAs, where a trusted planner generates a complete execution graph with conditional branches before any observation of potentially malicious content, providing provable control flow integrity guarantees against arbitrary instruction injections. Although this architectural isolation successfully prevents instruction injections, we show that additional measures are needed to prevent Branch Steering attacks, which manipulate UI elements to trigger unintended valid paths within the plan. We evaluate our design on OSWorld, and retain up to 57% of the performance of frontier models while improving performance for smaller open-source models by up to 19%, demonstrating that rigorous security and utility can coexist in CUAs.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes