agent/papers/items/2026-2604-10577-the-blind-spot-of-agent-safety-how-benign-user-instructions-expose-critical-vuln.md

Paper: The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents

---

type: paper title: "The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents" authors: Xuwei Ding, Skylar Zhai, Linxin Song, Jiate Li, Taiwei Shi, Nicholas Meade, Siva Reddy, Jian Kang, et al. year: 2026 venue: arXiv url: https://arxiv.org/abs/2604.10577 code_url: source: arxiv collected_at: 2026-07-08 published_at: 2026-04-12 updated_at: 2026-04-17 status: queued relevance: high topics:

• agent-evaluation
• agent-safety
• multi-agent
• rag
• tool-use methods:

benchmarks:

models:

datasets:

• cs.CR
• cs.AI related_concepts:

related_jobs:

related_experiments:

related_projects:

collection_score: 17 collection_queries: agent-safety

One-line Takeaway

Auto-collected from arXiv because it matched the Agent collection queries. Needs human skim.

Why Collected

• matched queries: agent-safety
• inferred topics: agent-evaluation, agent-safety, multi-agent, rag, tool-use
• arXiv categories: cs.CR, cs.AI
• collection score: 17

Review Checklist

• Does this paper directly inform Agent architecture, evaluation, memory, tools, safety, coding agents, GUI/browser agents, or multi-agent workflows?
• Does it include a benchmark, dataset, code, or reproducible experimental setup?
• Should it be promoted from queued to skimmed or summarized?

Links

• arXiv: https://arxiv.org/abs/2604.10577