mlsanityData to 14 Oct 2026

ARMOR: Agentic Reasoning for Methods Orchestration and Reparameterization for Robust Adversarial Attacks

arXiv 2601.18386 ↗First posted 26 Jan 2026cs.CVcited by 1

On the map: LLM reasoning → Reasoning for agents, planning and code

also in 1 more
Abstract (arXiv)

Existing automated attack suites operate as static ensembles with fixed sequences, lacking strategic adaptation and semantic awareness. This paper introduces the Agentic Reasoning for Methods Orchestration and Reparameterization (ARMOR) framework to address these limitations. ARMOR orchestrates three canonical adversarial primitives, Carlini-Wagner (CW), Jacobian-based Saliency Map Attack (JSMA), and Spatially Transformed Attacks (STA) via Vision Language Models (VLM)-guided agents that collaboratively generate and synthesize perturbations through a shared ``Mixing Desk". Large Language Models (LLMs) adaptively tune and reparameterize parallel attack agents in a real-time, closed-loop system that exploits image-specific semantic vulnerabilities. On standard benchmarks, ARMOR achieves improved cross-architecture transfer and reliably fools both settings, delivering a blended output for blind targets and selecting the best attack or blended attacks for white-box targets using a confidence-and-SSIM score.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 32 of its 36 references that are in the map. First posted on arXiv 26 Jan 2026.

Builds on

32 of 36 references.

References · 32

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  4. 2016Densely Connected Convolutional Networkscited by 6,720
  5. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  6. 2013Intriguing properties of neural networkscited by 5,155
  7. 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,568
  8. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
  9. 2023A Survey on Large Language Model based Autonomous Agentscited by 2,242
  10. 2023Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyondcited by 2,224
  11. 2023AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversationcited by 1,988
  12. 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,451
20 more
  1. 2019FaceForensics++: Learning to Detect Manipulated Facial Imagescited by 1,101
  2. 2016Practical Black-Box Attacks against Machine Learningcited by 903
  3. 2015The Limitations of Deep Learning in Adversarial Settingscited by 790
  4. 2019Square Attack: a query-efficient black-box adversarial attack via random searchcited by 685
  5. 2018Improving Transferability of Adversarial Examples with Input Diversitycited by 599
  6. 2016Delving into Transferable Adversarial Examples and Black-box Attackscited by 536
  7. 2018Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Surveycited by 460
  8. 2019Evading Defenses to Transferable Adversarial Examples by Translation-Invariant Attackscited by 442
  9. 2025Multi-Agent Collaboration Mechanisms: A Survey of LLMscited by 431
  10. 2019Nesterov Accelerated Gradient and Scale Invariance for Adversarial Attackscited by 355
  11. 2019Simple Black-box Adversarial Attackscited by 251
  12. 2020The Creation and Detection of Deepfakes: A Surveycited by 247
  13. 2020DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detectioncited by 233
  14. 2018Spatially Transformed Adversarial Examplescited by 195
  15. 2024Large Language Model Agent for Hyper-Parameter Optimizationcited by 64
  16. 2018Adversarial Examples - A Complete Characterisation of the Phenomenoncited by 11
  17. 2024Efficient Black-box Adversarial Attacks via Bayesian Optimization Guided by a Function Priorcited by 11
  18. 2025WILD: a new in-the-Wild Image Linkage Dataset for synthetic image attributioncited by 9
  19. 2025Adversarial Agents: Black-Box Evasion Attacks with Reinforcement Learningcited by 3
  20. 2025MS-GAGA: Metric-Selective Guided Adversarial Generation Attackcited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026ARMOR++: Agentic Orchestration of a Multi-Domain Primitive Set for Transferable Attacks on Deepfake Detectorscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025ARMs: Adaptive Red-Teaming Agent against Multimodal Models with Plug-and-Play Attacks
  2. 2025Agentic Moderation: Multi-Agent Design for Safer Vision-Language Models
  3. 2025AegisLLM: Scaling Agentic Systems for Self-Reflective Defense in LLM Security
  4. 2026Red-teaming the Multimodal Reasoning: Jailbreaking Vision-Language Models via Cross-modal Entanglement Attacks
  5. 2026SABER: A Stealthy Agentic Black-Box Attack Framework for Vision-Language-Action Models
  6. 2026VALD: Multi-Stage Vision Attack Detection for Efficient LVLM Defense
  7. 2025EVA: Evolving Semantic Adversaries for Red-Teaming GUI Agents Against Environmental Injection Attacks
  8. 2026RIFT-Bench: Dynamic Red-teaming For Agentic AI Systems