Resources & Knowledge Base
Research papers, technical whitepapers, and practical guides from our work in security, automation, blockchain, and fullstack engineering.
Whitepapers & Research
In-depth research from our security, AI, and automation work, including the Koreshield threat intelligence framework
LLM-Firewall: A Lightweight Middleware Architecture for Real-Time Detection of Indirect Prompt Injection in Enterprise RAG Systems
Presents a middleware architecture for real-time detection of indirect prompt injection in enterprise RAG systems. Introduces a cross-document correlation engine that identifies fragmented attack payloads distributed across multiple retrieved chunks. Evaluated across 5,000 adversarial scenarios.
LLM-Firewall: A Novel Taxonomy of Indirect Prompt Injection Attacks in Enterprise RAG Systems
Introduces a five-dimensional taxonomy for classifying indirect prompt injection attacks against enterprise RAG deployments, covering injection vector, operational target, persistence mechanism, enterprise context, and detection complexity.
Tool Call Hijacking in Agentic Pipelines: A Practical Demonstration
Demonstrates attacks where malicious instructions in tool outputs redirect autonomous LLM agents. Proof-of-concept across LangChain, AutoGen, and Claude Tools confirms the weakness is architectural rather than framework-specific.
Technical Guides
Practical development guides drawn from our project work
Setting Up a Production-Ready Node.js API
Step-by-step guide covering project structure, authentication, error handling, logging, rate limiting, and deployment best practices.
React Component Architecture for Scale
Patterns for organizing React applications: compound components, render props, custom hooks, and state management strategies that scale with your team.
Blockchain Development Fundamentals
From Solidity basics to deploying and verifying smart contracts on Ethereum and Polygon. Includes testing with Hardhat and security considerations.
Need More Information?
Can't find what you're looking for? Contact us and we'll be happy to help.