AI security
Agentic AI, Model Context Protocol and LLM security: how autonomous systems get attacked, and how to govern them before they act.
Papers
Post-Quantum Security Framework for Model Context Protocol Using ContextGuard
Sunil Gentyala, P. Krishna Sunil, Vamshi Lande, Akhila Kasturi, Suresh Kumar Darisi, Ruhisulthana Shaik
Agentic Security Validation Framework for Retrieval-Augmented and Tool-Enabled Large Language Model Systems (ARGUS)
Sunil Gentyala, N. Sudhakar Reddy, Kothapalli Chaitanya Deepthi, Pathapati Saroja, Kammara Venkatarangaiah Achari, Ruhisulthana Shaik
A Zero-Trust Supply Chain Security Framework for Model Context Protocol-Based AI Systems
Sunil Gentyala, Ch. Srinivas, Raghu Dhumpati
A Unified Mathematical Framework for Secure Multi-Agent Large Language Model Pipelines (TRACE-MAS)
Sunil Gentyala, Y. Geetha Reddy, Manasa Pendyala, Vishnu Gatla, Sundarigari Manoj, Ruhisulthana Shaik
CyberFuse-CI: Adversarially Resilient Vulnerability Detection Through Heterogeneous Multi-Source Data Fusion and LLM-Augmented Reasoning
Sunil Gentyala, Sunil Kumar Mudusu, Satish Kumar Allani, Manasa Pendyala, Rakesh Prakash, Praveen Kumar Mannam
A Multi-Stage NLP Framework for Enterprise Data Protection in Public LLM Interactions
Sunil Gentyala, Nimmagari Tejasri, Sunil Kumar Mudusu
AdaptFlow: A Self-Optimizing AI-Driven Data Pipeline Architecture for Real-Time Inference at Scale
Sunil Kumar Mudusu, Sreepal Reddy Bolla, Sunil Gentyala
Zero-Trust Data Pipelines for AI Systems: A Framework for Secure, Verifiable, and Auditable Data Engineering
Sunil Kumar Mudusu, Sunil Gentyala
On Auditing MCP Deployments
Sunil Gentyala, Praveen Kumar Mannam
Articles and whitepapers
MCP Fixed Secret Handling. URL Elicitation Moved the Phishing Risk Into the Browser
When the Prompt Becomes the Payload: A Practical Pen-Testing Guide for GenAI, LLM and RAG Applications
Shadow Agents: How IT Leaders Must Govern Headless AI Before It Breaks the Enterprise
Securing the Swarm: Governance, Attack Surfaces, and Zero-Trust Architectures in Multi-Agent AI Environments
Who Authorized the Algorithm? Reckoning with Ungoverned AI
1,800+ MCP Servers Exposed Without Authentication: How Zero Trust Can Secure the AI Agent Revolution
LLM-Generated Passwords Are Indefensible. Your Codebase May Already Prove It
The Agentic Trust Deficit: Why MCP's Authentication Vacuum Demands a New Security Paradigm
MCP is the Backdoor Your Zero-Trust Architecture Forgot to Close
The Sentinel Intelligence: A CISO's Guide to Sovereign Security in the Age of AGI
Why 2025's Agentic AI Boom Is a CISO's Worst Nightmare
Securing the Knowledge Layer: Enterprise Security Architecture Frameworks for Proprietary Data Integration With Large Language Models
The Silent Saboteur: When AI Autoscaling Goes Rogue
Navigating the Liminal Edge of AI Security: Deconstructing Prompt Injection, Model Poisoning, and Adversarial Perturbations in the Cognitive Cyber Domain
Beyond Silos: How DDI-AI Integration Is Redefining Cyber Resilience
Securing Application-to-Application Traffic with AI/AGI/ML-Powered Virtual Firewalls
The New Frontier of Cyber Threats: Unpacking Prompt Injection, Model Poisoning and Adversarial Attacks in AI Security
Projects
AegisSwarm-Core
Zero-trust governance for autonomous multi-agent AI networks. Published by the Cloud Security Alliance and cited externally as a reference implementation.
GSH Framework
Governed Security Hunting: agentic AI threat hunting mapped to MITRE ATLAS and NIST CSF 2.0, with a full playbook set and CI.
AEGIS Academic Integrity Checker
Local-first MCP server and CLI that checks papers for hallucinated citations, AI-text and plagiarism signals, and venue guideline compliance. Listed on mcprush; on PyPI as aegis-integrity.
ContextGuard
Zero-trust middleware for Model Context Protocol security.
ContextGuard-MCP-PQC
Applies RFC 9794 PQ/T hybrid terminology to MCP across four layers, hardening handshakes with hybrid ML-KEM and ML-DSA against harvest-now-decrypt-later and downgrade attacks.
ARGUS
Autonomous vulnerability discovery for retrieval-augmented and tool-enabled LLM systems.
genai-llm-pentest-harness
Reference implementation for the CSO Online pen-testing guide on GenAI, LLM and RAG applications.
PRAVAL
Ten-phase agentic vulnerability-assessment framework.
TRACE-MAS
Cryptographic framework securing multi-agent LLM pipelines against drift, spoofing and prompt injection.
CausalNetTwin
Causal digital-twin framework for network security reasoning.
CADLP
Context-aware DLP proxy that detects and redacts credentials, PII and proprietary code before prompts reach external LLM APIs.
LaptopAI-Agent
Fully local, privacy-first autonomous agent built on LangGraph, Ollama, ChromaDB RAG and MCP tools.