Security Intelligence Framework: Vulnerability Detection Dataset and Reproducibility Package

Comprehensive dataset and reproducibility package for the Security  Intelligence Framework research presented in IEEE TDSC submission  \Security Intelligence Framework: A Unified Mathematical Approach for  Autonomous Vulnerability Detection\.   This package includes:  - 50,000+ labeled vulnerability samples across multiple programming  languages  - Real CVE case studies (Log4j, Heartbleed, Struts2, Citrix ADC,  Zerologon)  - Complete reproducibility environment with Docker containerization  - Evaluation datasets for statistical validation  - Synthetic vulnerability generation templates  - Configuration files and documentation   The dataset enables reproduction of all experimental results presented in  the manuscript, including the 98.5% precision and 97.1% recall performance   metrics, statistical significance testing, and real-world validation on  12.35 million lines of production code.