Rui Yang

Rui Yang

Security Informatics Graduate Student

Johns Hopkins University

Graduate student in Security Informatics at Johns Hopkins University. I am fortunate to be mentored by Prof. Yinzhi Cao in web security research, which I am deeply passionate about. Previously graduated with a Computer Science degree from East China Normal University (ECNU) as an Outstanding Graduate.

My research focuses on Host Name Pollution vulnerabilities, and I serve as a Teaching Assistant for Web Security. I was also fortunate to work as a Summer 2025 Research Assistant at University of Texas at Dallas under the guidance of Prof. Wei Yang, where I gained valuable exposure to AI security.

Discovered 10+ CVEs, with CVE-2025-43930–43933 publicly disclosed. Preparing submissions for IEEE S&P 2026, USENIX Security 2026, and ACM ASIACCS 2026.

ryang54@jh.eduAlso known as Brooke

Research

Working on research in web security and AI security, learning and contributing to the field through systematic vulnerability analysis and responsible disclosure practices.

1

Host Name Pollution

Conducting first-author research on systematic detection and large-scale measurement of host header injection vulnerabilities. Framework analysis has led to the discovery of CVE-2025-43930–43933.

IEEE S&P 2026First Author
2

Web Security Research

Working on comprehensive analysis of web application vulnerabilities including prototype pollution, DOM clobbering, and script gadgets. Contributing to automated detection and patching framework development.

SafeLookupUSENIX Security
3

AI Security & MoE

Working on adversarial vulnerabilities in Mixture-of-Experts architectures and exploring defense mechanisms for large language models. Research focuses on efficiency degradation and robustness failure analysis.

MoEFailACM ASIACCS

Publications in Preparation

Contributing to the academic community through research publications targeting top-tier security conferences, with ongoing work across multiple research areas.

FIRST AUTHORIEEE S&P 2026

Poisoned by the Host: Systematic Detection and Large-Scale Measurement of Host Name Pollution in Web Frameworks

IEEE Symposium on Security and Privacy 2026 (Target)

📅 Submission Deadline: November 14, 2025 (Abstract due: November 7, 2025)

First-author research presenting a comprehensive study of Host Name Pollution (HNP) vulnerabilities across popular web frameworks. The work involves developing automated detection tools and conducting large-scale measurements, which have led to the discovery of CVE-2025-43930–43933 and additional security issues.

Host Name PollutionCVE DiscoveryFramework AnalysisLarge-Scale Study

SafeLookup: Automated Detection and Patching of Prototype Pollution, DOM Clobbering, and Script Gadget Vulnerabilities

USENIX Security 2026 (Target)

📅 Submission Deadline: February 6, 2026 (Abstract due: January 30, 2026)

Developing automated detection and patching systems for client-side JavaScript vulnerabilities using machine learning and static analysis techniques. This research aims to address security gaps in modern web applications through comprehensive gadget chain analysis.

JavaScript SecurityMachine LearningAutomated Patching

MoEFail: Triggering Efficiency Degradation and Robustness Failures in Mixture-of-Experts Systems

ACM ASIACCS 2026 (Target)

📅 Submission Deadline: December 13, 2025

Investigating adversarial vulnerabilities in Mixture-of-Experts architectures and exploring defense mechanisms for large language models. This research contributes to understanding security concerns in modern AI systems.

AI SecurityAdversarial MLLLM Security

CVE Discoveries

Contributing to cybersecurity through systematic vulnerability research, with 10+ CVEs discovered and multiple CVEs publicly disclosed through responsible disclosure practices.

CVE

Published CVEs

CVE-2025-43930 – CVE-2025-43933

Publicly disclosed host header injection vulnerabilities in popular web frameworks. These findings represent a subset of the 10+ CVEs identified through systematic security analysis and automated detection methodologies.

Host Name PollutionPublicly DisclosedWeb Frameworks
🔍

Total CVE Discoveries

10+ CVEs Discovered

Comprehensive vulnerability research across multiple web frameworks and applications. While CVE-2025-43930–43933 have been publicly disclosed, additional CVEs are currently undergoing responsible disclosure processes with relevant vendors and organizations.

10+ Total CVEsResponsible DisclosureOngoing Research

Experience

Building expertise through research, internships, and academic collaborations.

Johns Hopkins University

Master of Science in Security Informatics

September 2024 - July 2026

Graduate student specializing in web security and AI security research. Conducting cutting-edge research in vulnerability discovery, host header injection analysis, and machine learning robustness with multiple top-tier conference publications in preparation.

Web SecurityAI SecurityResearch

East China Normal University

Bachelor of Science in Computer Science and Technology

September 2020 - July 2024 | Shanghai, China

Completed undergraduate studies at East China Normal University, a prestigious 985, 211, and "Double First-Class" research university in China. Focused on computer science fundamentals, software engineering, and security. Graduated as Outstanding Graduate (top 10%) with multiple academic excellence scholarships and leadership roles in student organizations.

Outstanding GraduateAcademic ExcellenceLeadership

Host Name Pollution Research

Independent Researcher | Johns Hopkins University

February 2025 - Present | Supervisor: Prof. Yinzhi Cao

First-author research on systematic detection and large-scale measurement of Host Name Pollution vulnerabilities, conducted under the guidance of Prof. Yinzhi Cao. Designed and executed comprehensive black-box and gray-box audits to uncover trust-model flaws in email, password-reset, and token-binding workflows. Developed white-box analyzers with taint tracking for Python/JavaScript/Java/PHP/Ruby/Go/Rust frameworks.

First AuthorIEEE S&P 2026CVE Discovery

DOM Clobbering, Prototype Pollution, and Script Gadgets Research

Research Assistant | Johns Hopkins University

November 2024 - Present | Supervisor: Prof. Yinzhi Cao

Conducting large-scale vulnerability research on prototype pollution, DOM clobbering, and script gadgets in JavaScript frameworks under Prof. Yinzhi Cao's supervision. Developing automated pipeline for gadget chain extraction and validation, with ML-based classifiers for high-impact candidate prioritization and LLM-based automatic repair framework.

SafeLookupUSENIX SecurityAuto-Patching

MoE Robustness Research

Summer Research Assistant | University of Texas at Dallas

Summer 2025 | Supervisor: Prof. Wei Yang

Conducted summer research with Prof. Wei Yang at University of Texas at Dallas, gaining valuable exposure to AI security research. Investigated robustness and efficiency limitations in dynamic Mixture-of-Experts (MoE) systems under real-world input variations. Developed perturbation-based attacks to systematically expose routing imbalance and efficiency degradation, with integrated measurement pipeline for latency, FLOPs, and expert activation analysis.

MoE AnalysisACM ASIACCSAI Security

Web Security Teaching Assistant

Teaching Assistant | Johns Hopkins University

Fall 2025 - Present | Instructor: Prof. Yinzhi Cao

Honored to serve as Teaching Assistant for Prof. Yinzhi Cao's Web Security course. Assisting students with web security concepts, vulnerability analysis, and hands-on security exercises. Contributing to curriculum development and providing guidance on practical security research methodologies under Prof. Cao's mentorship.

Teaching AssistantWeb SecurityCurriculum DevelopmentStudent Mentoring

Internship Experience

Professional experience in enterprise security development and AI/LLM security applications.

🏦

Java Full-Stack Developer

Capgemini Shanghai, China

July 2023 - February 2024

Built and maintained secure backend modules for HSBC's eCommerce and payment systems using Spring Boot and Vue.js. Implemented enterprise-grade web security controls including JWT authentication, session lifecycle hardening, CSRF defenses, and role-based access control (RBAC).

Spring BootVue.jsJWT AuthRBAC
🏥

Web & LLM Security

Medivoice Baltimore, US

October 2024 - March 2025

Developed and secured an LLM-based appointment scheduling system by integrating OpenAI APIs with sensitive healthcare web services. Designed tests and protective measures to ensure LLM output consistency and block metadata injection or unauthorized API calls.

LLM SecurityOpenAI APIHealthcare SecurityAPI Security

Undergraduate Research

Innovative projects and research during undergraduate studies at East China Normal University, a prestigious 985, 211, and "Double First-Class" research university in Shanghai.

💳

BiFu: Aggregated Payment Platform

Independent Researcher | ECNU

January 2024 - June 2024 | Supervisor: Prof. Jian Jin

Designed and deployed a production-ready aggregated payment platform using Spring Boot, Spring Cloud, and Vue.js, integrating WeChat, Alipay, and UnionPay. Built a secure microservice backend with centralized authentication (OAuth2, JWT) and robust multi-tenant login for merchants and admins.

Payment SystemsMicroservicesOAuth2/JWTMulti-tenant

Online Debug (EOD) Platform

Team Leader | ECNU

August 2022 - September 2023 | Supervisors: Asso. Prof. Ailian Fang & Asso. Prof. Fei Xu

Launched ECNU's first online programming assistant with live Q&A, algorithm sharing, and peer code review. Designed and built the entire tech stack, including forum logic, authentication and secure role-based access control. Selected as Shanghai University Student Innovative Projects.

Team LeadershipFull-Stack DevelopmentInnovation ProjectRBAC
⛓️

Xing Lian Blockchain Project

Team Leader | ECNU

August 2022 - August 2023 | Supervisors: Prof. Aimin Zhou & Kaijing Cai (CIO, East Money Information Co.)

Designed a dual-chain blockchain architecture enabling secure, decentralized data exchange between business entities. Developed a novel threshold signature-based authorization protocol, achieving 42.1% time cost reduction. Won Gold Prize in "Challenge Cup" National College Student Entrepreneurship Plan Competition.

BlockchainThreshold SignaturesGold Prize WinnerEnterprise Collaboration

Skills & Tools

Technical expertise and tools for cybersecurity research and development.

Web Security Research

Host Name PollutionPrototype PollutionDOM ClobberingScript Gadgets
🔧

Programming & Development

PythonJavaScriptJavaGoShellSQL

Security Tools & Frameworks

HHIHunterCodeQLJoernTaint TrackingSpring BootVue.jsOAuth2/JWTRBAC

AI/ML & MoE Analysis

MoE AnalysisAdversarial MLLLM SecurityOpenAI APIToken FuzzingRouting Analysis

Automated Analysis

Dynamic AnalysisStatic AnalysisFuzzingSymbolic DOMGadget ChainsAuto-Patching
📊

Research & Development

Large-Scale StudiesFramework AnalysisBlockchainMicroservicesPayment SystemsMulti-tenant

Awards & Recognition

Academic excellence, research contributions, and recognition for innovation and leadership.

Challenge Cup Gold Prize

National College Student Entrepreneurship Plan Competition

Gold Prize winner for the Xing Lian blockchain project, demonstrating exceptional innovation in dual-chain architecture and threshold signature protocols.

🥇

First Prize - Innovation Competition

Shanghai Female College Student Innovation and Entrepreneurship Competition (0.75%)

First Prize winner among top 0.75% of participants, recognizing outstanding innovation and entrepreneurship capabilities.

Outstanding Graduate

East China Normal University (10%)

Recognized as Outstanding Graduate among top 10% of students at East China Normal University, a prestigious 985, 211, and "Double First-Class" research university, acknowledging academic excellence and comprehensive achievements.

💰

Academic Excellence Scholarships

Multiple Corporate & University Scholarships

Recipient of Academic Excellence Scholarships from Nezha Technology Co., Ltd. (2%) and People's Financial Holdings Group (1%), recognizing outstanding academic performance.

CVE Discovery Recognition

10+ CVE Assignments

Recognized for discovering and responsibly disclosing 10+ security vulnerabilities in popular web frameworks, with CVE-2025-43930–43933 publicly disclosed.

👥

Leadership Recognition

Student Union President & Social Practice

President of School of Computer Science and Technology Students' Union, and Social Practice Outstanding Individual (0.6%), demonstrating leadership and community engagement.

Contact

Let's connect and explore opportunities for collaboration.

ryang54@jh.eduJohns Hopkins University