Front-End Developer
Security Enthusiast
Research Scholar
MCP Builder
Stay safe from cyber
MCP Pioneer.
Security Researcher.
I build MCP servers for AI agents — and publish the research to make them secure.
research
published
PROJECTS
COMPLETED
Who Am I?
I build secure systems.
I break the ones that aren’t.
AI Security Researcher & Front-End Engineer — working at the intersection of threat intelligence and immersive web interfaces.
I’m Mahipal — a cybersecurity researcher and front-end developer with an M.Sc. in AI Security from a German university. I build tools that defend against the attacks I study — and interfaces precise enough to match.
My work spans EmailGuard AI (adversarial phishing detection), multi-agent AI systems, MCP servers, and production-grade apps in React and Next.js — every one of them threat-modeled, OWASP-aware, and built to last.
Research
Peer-reviewed work on the edges of AI, security, and human behaviour
An Enhanced Haze Removal: Using DCP and Enriched-Invariant Features
Key finding: Enhanced Dark Channel Prior with Enriched-Invariant Features improves haze removal accuracy in severely degraded scenes beyond existing methods.
Developed and implemented a MATLAB-based algorithm combining Dark Channel Prior (DCP) and Enriched-Invariant Features for image dehazing. Peer-reviewed and published in Lecture Notes in Networks and Systems, Springer.
Read Paper →Cybersecurity Awareness Among Young Adults: An Analytical Study
Key finding: 75% of young adults use 2FA but only 15% understand encryption — confidence in data protection (R² = 0.46) is the strongest predictor of secure behaviour.
Survey of 68 young adults (18–27) across Germany and India assessing phishing recognition, 2FA adoption, and encryption literacy. Presented at Electronic Imaging 2025, San Francisco. Recommends gamification-based training.
Read Paper →MCP Tools
Building the infrastructure layer for AI agents
Model Context Protocol (MCP) is the open standard that lets AI assistants connect to external tools and systems. Adopted by Google, Microsoft, and IBM, MCP has become the backbone of AI agent infrastructure in 2025. I've been building MCP servers since early adoption — not using the tools, building them.
CVE MCP Server
Production-grade security intelligence for Claude. Real-time CVE lookup, EPSS scoring, MITRE ATT&CK mapping, Shodan, VirusTotal, threat intel, and risk scoring — all through natural language.
PostgreSQL MCP Server
AI-powered PostgreSQL management and monitoring. Automated database operations, diagnostics, performance analysis, security auditing, and optimization — all through natural language with Claude and ChatGPT.
GitHub →MCP Web Scrape
Transforms messy HTML into clean, agent-ready content. Markdown conversion, automatic citations, robots.txt compliance, ETag caching, and AI-powered summarization — any AI assistant, out of the box.
GitHub →MCP Windows Server
Lets Claude, ChatGPT, and other AI assistants control Windows through natural language. File operations, system control, web automation, and Office integration. Ask it to reorganize your Downloads. Watch it happen.
GitHub →Cybersecurity Skills
Structured skills for AI agents — MITRE ATT&CK mapped
Largest open-source skills collection for AI agents — deploys instantly to Claude Code, Copilot, Cursor, Gemini CLI, and 20+ more via the agentskills.io standard.
Latest Projects
Perfect solutions for digital experience
Claude Agent Dashboard - Real-Time Agent Team Monitor
Stop flying blind when running multi-agent workflows. Claude Agent Dashboard gives X-ray visibility into Claude Code agent teams — live task status, inter-agent message tracing, WebSocket streaming, and system metrics, all from a single unified dashboard. Built with JavaScript, Vite, and Node.js with zero-config setup.
GitHub LinkSK Studio - Fine Art Photography Portfolio
SK Studio & Co. is a modern, minimalist photography portfolio website built for a fine art photography business specializing in wedding and portrait photography. Built with React 18.3.1 and Vite, it features category-based galleries, lightbox viewer with keyboard navigation, comprehensive contact forms with validation, and optimized image loading. The project showcases professional photography work with an elegant, user-focused design that captures timeless moments with artistry and care.
View Live SiteMCP Web Scrape - Intelligent Web Content Extraction
"MCP Web Scrape" is a TypeScript-based web scraping framework that transforms messy HTML into clean, agent-ready content. Built on the Model Context Protocol (MCP), it provides intelligent content extraction with features including Markdown conversion, automatic citations, robots.txt compliance, ETag caching, and AI-powered summarization for seamless integration with AI assistants.
MCP Windows Server - AI-Powered Windows Automation
"MCP Windows Server" is a Python-based automation framework that enables AI assistants like Claude and ChatGPT to control Windows systems through natural language commands. This server implements the Model Context Protocol (MCP) to provide secure, comprehensive system-level automation with 200+ tools for file operations, system control, web automation, and Office integration.
An Enhanced Haze Removal: Using DCP and Enriched-Invariant Features
The project proposes an enhanced method for removing haze from images using Dark Channel Prior (DCP) and Enriched-Invariant Features. This approach aims to improve the accuracy and efficiency of existing haze removal techniques, particularly in situations where the visibility is severely impaired by haze.
Read PublicationEmailGuard - Proactive Email Phishing Detection using Artificial Intelligence
Developed and implemented "EmailGuard," an AI-driven system for detecting phishing emails using 50 extensive features. The system leverages machine learning algorithms and natural language processing to analyze email content and metadata, distinguishing between legitimate and phishing emails. This project significantly enhanced cybersecurity measures, reduced phishing incidents, and improved the overall security posture of the organization.
Multicast Wireless Routing Protocol In Java Programming
The Multicast Wireless Routing Protocol project in Java programming aims to develop a protocol for efficient data transmission over wireless networks using multicasting. The project involves implementing algorithms to optimize data transmission and minimize delays and packet losses.
Vampire Attacks Draining Life from Wireless Ad Hoc Sensor Networks
"Vampire Attacks" is a Java-based application that simulates and analyzes the impact of vampire attacks on wireless sensor networks. This application explores the vulnerabilities of such networks and mimics the behavior of energy-draining malicious nodes.
Read MorePack India Packers and Movers Website Development Project
I successfully crafted a professional website for Pack India Packer and Movers, highlighting their top-tier relocation services. The site provides seamless user experience, valuable information, and easy client interaction, bolstering Pack India's online presence and their mission of smooth moving solutions.
View Live SiteSkills
Technical expertise and tools
Programming
Web Development
AI & Machine Learning
Cybersecurity
Cloud & DevOps
MCP & AI Tools
