ADeptS-Bench: Measuring the Trustworthiness of Computer Use Agents Across Devices

arXiv:2608.26204 · cs.CR, cs.AI, cs.SE · Submitted 2026-08-25 · Read on arXiv

cs.CR, cs.AI, cs.SE

Submitted: 2026-08-25

Updated: 2026-09-15

Code: https://github.com/facebookresearch/adepts-bench

Terminology

• Benchmark — 54× in this paper · explained in 2 episodes
• Grounding — 18× in this paper · explained in 3 episodes
• LLM judge — 13× in this paper · explained in 2 episodes
• Calibration — 11× in this paper · explained in 7 episodes
• Ablation — 8× in this paper · explained in 1 episode
• Benchmarking — 8× in this paper · explained in 3 episodes
• Core — 5× in this paper · explained in 2 episodes
• False Refusal Rate — 5× in this paper
• Hallucination — 5× in this paper · explained in 3 episodes
• Accuracy — 4× in this paper · explained in 1 episode
• Affordance — 4× in this paper · explained in 1 episode
• Attack Success Rate — 4× in this paper
• Task Success Rate — 4× in this paper
• Visual Grounding — 3× in this paper · explained in 2 episodes
• Adaptation — 2× in this paper · explained in 1 episode
• Attack Success Rate (ASR) — 2× in this paper · explained in 2 episodes
• Ablation study — 1× in this paper · explained in 1 episode
• Adversarial Attacks — 1× in this paper · explained in 1 episode
• Adversarial Vulnerabilities — 1× in this paper · explained in 1 episode
• Alignment — 1× in this paper · explained in 1 episode
• Annotator Agreement — 1× in this paper · explained in 1 episode
• Contextual Understanding — 1× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Scaffolding — 1× in this paper · explained in 3 episodes

Sources

• Constitutional AI: Harmlessness from AI Feedback
• KnowU-Bench: Towards Interactive, Proactive, and Personalized Mobile Agent Evaluation
• Navi-plus: Managing Ambiguous GUI Navigation Tasks with Follow-up Questions
• The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents
• ADEPTS: A Capability Framework for Human-Centered Agent Design
• Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned
• Code Agent can be an End-to-end System Hacker: Benchmarking Real-world Threats of Computer-use Agent
• Qwen3-VL Technical Report
• AmbiBench: Benchmarking Mobile GUI Agents Beyond One-Shot Instructions in the Wild
• DigiData: Training and Evaluating General-Purpose Mobile Control Agents

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs