# New Papers Show Images and Retrieved Documents Can Hijack Agents That Now Hold Payment Credentials

One framework tests whether a visual patch on screen produces verifiable consequences in a computer-use agent's environment, and a second measures how retrieval-augmented systems repeat poisoned text.

- Published: 2026-09-10T06:24:02.019Z
- Canonical: https://polylog.news/ai/2026-09-10/new-papers-show-images-and-retrieved-documents-can-hijack-ag
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [arXiv cs.CR](https://arxiv.org/abs/2609.09212), [arXiv cs.CR](https://arxiv.org/abs/2609.09243), [arXiv cs.CR](https://arxiv.org/abs/2609.09229)

Two security papers posted this week address the exact failure mode that consumer and enterprise agents introduce. AgentHijack sets out an end-to-end evaluation framework for image-triggered command injection against computer-use agents, an…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-09-10/new-papers-show-images-and-retrieved-documents-can-hijack-ag (subscription information: https://polylog.news/pricing).