mlsanityData to 7 Oct 2026

MaskClaw: Edge-Side Personalized Privacy Arbitration for GUI Agents with Behavior-Driven Skill Evolution

arXiv 2605.28646 ↗First posted 27 May 2026cs.CR cs.CLcited by 1

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

GUI agents rely on screenshots to infer intent and operate across applications, but these screenshots often contain private messages, medical records, payment credentials, and workplace-specific workflows. Privacy decisions in this setting depend on task, recipient, application state, and user role, yet static PII detectors miss these boundaries and cloud-side VLM reasoning can upload the raw screen before deciding what should be protected. We present MaskClaw, an edge-side privacy arbitrator for GUI agents. MaskClaw extracts local visual evidence, retrieves user- and task-specific policy memory, and decides Allow, Mask, or Ask before raw screenshots leave a trusted user- or organization-controlled environment. In five designed skill-evolution scenarios, it turns corrections, cancellations, and edits into reusable privacy skills checked by a sandbox gate. We introduce P-GUI-Evo, a benchmark built from real UI patterns, reconstructed HTML screens, and sanitized labels. Experiments show that pattern matching, cloud reasoning, and routing alone tend to over-confirm, over-mask, or expose raw screenshots under the same protocol. The artifact is available at https://github.com/Theodora-Y/MaskClaw.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 25 of its 41 references that are in the map. First posted on arXiv 27 May 2026.

Builds on

25 of 41 references.

Closest ancestors · 2

Same direction, newest first

  1. 2026MemPrivacy: Privacy-Preserving Personalized Memory Management for Edge-Cloud Agents10 May
  2. 2025Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents22 Sep

Other references · 23

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  2. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  3. 2020Extracting Training Data from Large Language Modelscited by 2,144
  4. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,800
  5. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
  6. 2024GPT-4V(ision) is a Generalist Web Agent, if Groundedcited by 547
  7. 2023AppAgent: Multimodal Agents as Smartphone Userscited by 407
  8. 2024Mobile-Agent: Autonomous Multi-Modal Mobile Device Agent with Visual Perceptioncited by 283
  9. 2026SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learningcited by 252
  10. 2024WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learningcited by 243
  11. 2024OmniParser for Pure Vision Based GUI Agentcited by 193
  12. 2024Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMscited by 173
11 more
  1. 2024GUI Agents with Foundation Models: A Comprehensive Surveycited by 110
  2. 2024AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agentscited by 105
  3. 2026SkillClaw: Let Skills Evolve Collectively with Agentic Evolvercited by 76
  4. 2026Memento-Skills: Let Agents Design Agentscited by 56
  5. 2024PAPILLON: Privacy Preservation from Internet-based and Local Language Model Ensemblescited by 54
  6. 2025PIG: Privacy Jailbreak Attack on LLMs via Gradient-based Iterative In-Context Optimizationcited by 10
  7. 2025Computer-Use Agents as Judges for Generative User Interfacecited by 8
  8. 2026Anonymization-Enhanced Privacy Protection for Mobile GUI Agents: Available but Invisiblecited by 5
  9. 2026TreeCUA: Efficiently Scaling GUI Automation with Tree-Structured Verifiable Evolutioncited by 4
  10. 2026VisualLeakBench: Auditing the Fragility of Large Vision-Language Models against PII Leakage and Social Engineeringcited by 4
  11. 2026Know Thy Enemy: Securing LLMs Against Prompt Injection via Diverse Data Synthesis and Instruction-Level Chain-of-Thought Learningcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026CAPED: Context-Aware Privacy Exposure Defense for Mobile GUI Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026GUIGuard-Bench: Toward a General Evaluation for Privacy-Preserving GUI Agents
  2. 2026PriMobiBench: Characterizing Visual Privacy Leakage in VLM-Driven Mobile GUI Agents
  3. 2026ASLEval: Measuring Privacy Exposure Displacement in LLM Agent Sessions
  4. 2026Mobile GUI Agent Privacy Personalization with Trajectory Induced Preference Optimization
  5. 2026WebPII: Benchmarking Visual PII Detection for Computer-Use Agents
  6. 2026PrivacySkills: How Privacy Guidance Shapes Source Selection in LLM Agents
  7. 2026AgentSCOPE: Evaluating Contextual Privacy Across Agentic Workflows
  8. 2026Minim: Privacy-Aware Minimal View for Agents via Trusted Local Sanitization