agent/papers/items/2026-2606-07595-visualleakbench-reproducible-action-boundary-propagation-failures-in-vision-lang.md

Paper: VisualLeakBench: Reproducible Action-Boundary Propagation Failures in Vision-Language Agents

---

type: paper title: "VisualLeakBench: Reproducible Action-Boundary Propagation Failures in Vision-Language Agents" authors: Youting Wang, Yuan Tang, Yitian Qian, Chen Zhao year: 2026 venue: arXiv url: https://arxiv.org/abs/2606.07595 code_url: source: arxiv collected_at: 2026-07-08 published_at: 2026-05-29 updated_at: 2026-05-29 status: queued relevance: high topics:

• agent-evaluation
• agent-safety
• memory
• tool-use
• workflow-agent methods:

benchmarks:

models:

datasets:

• cs.CV
• cs.AI
• cs.IR related_concepts:

related_jobs:

related_experiments:

related_projects:

collection_score: 16 collection_queries: language-agent

One-line Takeaway

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

Why Collected

• matched queries: language-agent
• inferred topics: agent-evaluation, agent-safety, memory, tool-use, workflow-agent
• arXiv categories: cs.CV, cs.AI, cs.IR
• collection score: 16

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/2606.07595