AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

RedTeamCUA: Realistic Adversarial Testing of Computer-Use Agents in Hybrid Web-OS Environments

Zeyi Liao, Jaylen Jones, Linxi Jiang, Yuting Ning, Eric Fosler-Lussier, Yu Su, Zhiqiang Lin, Huan Sun

ICLR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: LLM jailbreak and injection defense · 37 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use Operator with built-in confirmation and safety checks instead of Claude 3.7 Sonnet CUA / GPT-4o scaffoldingOperator · RTC-Bench DecoupledASR (%): -58.62% ASRImproved
Disable Operator's confirmation and safety checks (inattentive supervision) (ablation)Operator · RTC-Bench Decoupled (avg)ASR (%): 7.57 -> 30.89Hurt
Use Claude 3.7 Sonnet CUA specialization instead of base Claude 3.7 SonnetClaude 3.7 Sonnet · RTC-Bench Decoupled (avg)ASR (%): 39.33 -> 42.93Hurt
Use Claude 3.5 Sonnet CUA specialization instead of base Claude 3.5 SonnetClaude 3.5 Sonnet · RTC-Bench Decoupled (avg)ASR (%): 41.37 -> 31.21Improved
Evaluate End2End instead of Decoupled Eval (agent must navigate to injection itself)Operator (w/o checks) · RTC-Bench 50 high-risk tasksASR (%): 94.0 -> 42.0Improved
Use more capable Claude 4.5 Opus CUA instead of Claude 3.7 Sonnet CUA (End2End)Claude CUAs · RTC-Bench End2End (50 tasks)ASR (%): 50 -> 83Hurt
Upgrade Claude Opus 4.5 CUA to Claude Opus 4.6 CUAClaude Opus CUA · RTC-Bench End2End (50 tasks)ASR (%): 83 -> 50Improved
Add defensive system prompt instructing injection detection (ablation)Claude 3.7 Sonnet CUA and Operator (w/o checks) · RTC-Bench 50 high-risk tasksASR (%): ASR remain near 50%No clear effect
Add LlamaFirewall or PromptArmor detectors (ablation)CUAs · RTC-Bench 50 high-risk tasksinjection detection rate: best variant detects only 30%No clear effect
Use Meta SecAlign model-level defense (ablation)Meta SecAlign · RTC-Bench 50 high-risk tasksASR: follows malicious instructions in about half of tasksNo clear effect
Move from Confidentiality to Integrity goals for Claude 3.5 Sonnet on OwnCloudClaude 3.5 Sonnet · OwnCloudASR (%): 0.00 -> 48.67Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy