AgentTypo: Adaptive Typographic Prompt Injection Attacks against Black-box Multimodal Agents

• URL: http://arxiv.org/abs/2510.04257v1
• Date: Sun, 05 Oct 2025 15:46:56 GMT
• Title: AgentTypo: Adaptive Typographic Prompt Injection Attacks against Black-box Multimodal Agents
• Abstract summary: We introduce AgentTypo, a framework that mounts adaptive typographic prompt injection by embedding optimized text into webpage images.<n>Our automatic typographic prompt injection (ATPI) algorithm maximizes prompt reconstruction by substituting captioners while minimizing human detectability via a stealth loss.<n>We also develop AgentTypo-pro, a multi-LLM system that iteratively refines injection prompts using evaluation feedback and retrieves successful past examples for continual learning.
• Score: 22.88469633141419
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Multimodal agents built on large vision-language models (LVLMs) are increasingly deployed in open-world settings but remain highly vulnerable to prompt injection, especially through visual inputs. We introduce AgentTypo, a black-box red-teaming framework that mounts adaptive typographic prompt injection by embedding optimized text into webpage images. Our automatic typographic prompt injection (ATPI) algorithm maximizes prompt reconstruction by substituting captioners while minimizing human detectability via a stealth loss, with a Tree-structured Parzen Estimator guiding black-box optimization over text placement, size, and color. To further enhance attack strength, we develop AgentTypo-pro, a multi-LLM system that iteratively refines injection prompts using evaluation feedback and retrieves successful past examples for continual learning. Effective prompts are abstracted into generalizable strategies and stored in a strategy repository, enabling progressive knowledge accumulation and reuse in future attacks. Experiments on the VWA-Adv benchmark across Classifieds, Shopping, and Reddit scenarios show that AgentTypo significantly outperforms the latest image-based attacks such as AgentAttack. On GPT-4o agents, our image-only attack raises the success rate from 0.23 to 0.45, with consistent results across GPT-4V, GPT-4o-mini, Gemini 1.5 Pro, and Claude 3 Opus. In image+text settings, AgentTypo achieves 0.68 ASR, also outperforming the latest baselines. Our findings reveal that AgentTypo poses a practical and potent threat to multimodal agents and highlight the urgent need for effective defense.

Related papers

• Automatic Red Teaming for Implicit Vulnerabilities of Text-to-Image Models [58.59008920408855]
  We propose Adversarial Probing for Implicit VulnErabilities (AdvPIE) to expose implicit vulnerabilities without requiring access to the parameters of target models.<n> AdvPIE adopts a policy agent to generate and refine implicit adversarial prompts based on the feedback from a judge agent.<n>Experiments on standard and safety-aligned T2I models show that AdvPIE1 effectively uncovers implicit vulnerabilities.
  arXiv  Detail & Related papers  (2026-09-05T13:45:09Z)
• SnapGuard: Lightweight Prompt Injection Detection for Screenshot-Based Web Agents [32.00656418129339]
  We propose SnapGuard, a lightweight yet accurate method for detecting prompt injection attacks from webpage screenshots.<n>We show that SnapGuard achieves an F1 score of 0.75, outperforming GPT-4o-prompt while being 8x faster (1.81s vs. 14.50s) and introducing no additional memory overhead.
  arXiv  Detail & Related papers  (2026-04-28T12:32:21Z)
• Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection [37.5890320718138]
  We propose a red-teaming setting that overlays safety-aligned and harmless UI elements onto screenshots to misdirect the agent's visual grounding.<n>Our method uses a modular Editor-Overlapper-Victim pipeline and an iterative search procedure that samples multiple candidate edits.<n>Our attacks improve attack success rate by up to 4.4x over random injection on the strongest victims.
  arXiv  Detail & Related papers  (2026-04-09T05:32:34Z)
• Image-based Prompt Injection: Hijacking Multimodal LLMs through Visually Embedded Adversarial Instructions [7.575333495550402]
  We study Image-based Prompt Injection (IPI), a black-box attack in which adversarial instructions are embedded into natural images to override model behavior.<n>Our end-to-end IPI pipeline incorporates segmentation-based region selection, adaptive font scaling, and background-aware rendering.
  arXiv  Detail & Related papers  (2026-03-04T01:58:08Z)
• Automating Agent Hijacking via Structural Template Injection [18.856564341900555]
  A critical threat to the Large Language Model (LLM) ecosystem, agent hijacking enables adversaries to manipulate execution by injecting malicious instructions into retrieved content.<n>We propose Phantom, an automated agent hijacking framework that targets the fundamental architectural mechanisms of LLM agents.<n>By injecting optimized structured templates into the retrieved context, we induce role confusion and cause the agent to misinterpret the injected content as legitimate user instructions or prior tool outputs.
  arXiv  Detail & Related papers  (2026-02-18T23:52:14Z)
• SkillJect: Automating Stealthy Skill-Based Prompt Injection for Coding Agents with Trace-Driven Closed-Loop Refinement [120.52289344734415]
  We propose an automated framework for stealthy prompt injection tailored to agent skills.<n>The framework forms a closed loop with three agents: an Attack Agent that synthesizes injection skills under explicit stealth constraints, a Code Agent that executes tasks using the injected skills and an Evaluate Agent that logs action traces.<n>Our method consistently achieves high attack success rates under realistic settings.
  arXiv  Detail & Related papers  (2026-02-15T16:09:48Z)
• SCOPE: Prompt Evolution for Enhancing Agent Effectiveness [53.75986399936395]
  Large Language Model (LLM) agents are increasingly deployed in environments that generate massive, dynamic contexts.<n>While agents have access to this context, their static prompts lack the mechanisms to manage it effectively.<n>We introduce textbfSCOPE (Self-evolving Context Optimization via Prompt Evolution)<n>We propose a Dual-Stream mechanism that balances tactical specificity (resolving immediate errors) with strategic generality (evolving long-term principles)
  arXiv  Detail & Related papers  (2025-12-17T12:25:05Z)
• Backdoor Attacks on Open Vocabulary Object Detectors via Multi-Modal Prompt Tuning [5.0734761482919115]
  Open-vocabulary object detectors (OVODs) unify vision and language to detect arbitrary object categories based on text prompts.<n>We conduct the first study of backdoor attacks on OVODs and reveal a new attack surface introduced by prompt tuning.
  arXiv  Detail & Related papers  (2025-11-16T19:05:31Z)
• Poison Once, Control Anywhere: Clean-Text Visual Backdoors in VLM-based Mobile Agents [54.35629963816521]
  This work introduces VIBMA, the first clean-text backdoor attack targeting VLM-based mobile agents.<n>The attack injects malicious behaviors into the model by modifying only the visual input.<n>We show that our attack achieves high success rates while preserving clean-task behavior.
  arXiv  Detail & Related papers  (2025-06-16T08:09:32Z)
• AgentVigil: Generic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents [54.29555239363013]
  We propose a generic black-box fuzzing framework, AgentVigil, to automatically discover and exploit indirect prompt injection vulnerabilities.<n>We evaluate AgentVigil on two public benchmarks, AgentDojo and VWA-adv, where it achieves 71% and 70% success rates against agents based on o3-mini and GPT-4o.<n>We apply our attacks in real-world environments, successfully misleading agents to navigate to arbitrary URLs, including malicious sites.
  arXiv  Detail & Related papers  (2025-05-09T07:40:17Z)
• Manipulating Multimodal Agents via Cross-Modal Prompt Injection [34.35145839873915]
  We identify a critical yet previously overlooked security vulnerability in multimodal agents.<n>We propose CrossInject, a novel attack framework in which attackers embed adversarial perturbations across multiple modalities.<n>Our method outperforms state-of-the-art attacks, achieving at least a +30.1% increase in attack success rates.
  arXiv  Detail & Related papers  (2025-04-19T16:28:03Z)
• Dissecting Adversarial Robustness of Multimodal LM Agents [70.2077308846307]
  We manually create 200 targeted adversarial tasks and evaluation scripts in a realistic threat model on top of VisualWebArena.<n>We find that we can successfully break latest agents that use black-box frontier LMs, including those that perform reflection and tree search.<n>We also use ARE to rigorously evaluate how the robustness changes as new components are added.
  arXiv  Detail & Related papers  (2024-06-18T17:32:48Z)
• White-box Multimodal Jailbreaks Against Large Vision-Language Models [61.97578116584653]
  We propose a more comprehensive strategy that jointly attacks both text and image modalities to exploit a broader spectrum of vulnerability within Large Vision-Language Models. Our attack method begins by optimizing an adversarial image prefix from random noise to generate diverse harmful responses in the absence of text input. An adversarial text suffix is integrated and co-optimized with the adversarial image prefix to maximize the probability of eliciting affirmative responses to various harmful instructions.
  arXiv  Detail & Related papers  (2024-05-28T07:13:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.