I already have 11 relevant papers on cross-dataset malware detection, but I need a PRISMA-compliant systematic literature review that expands this set to about 30 peer-reviewed works published in 2024–2025. The review must clearly present research questions, search strategy, inclusion/exclusion criteria, quality appraisal, synthesis, and a PRISMA flow diagram. Alongside the review, I need a fully reproducible Python pipeline that trains and tests LightGBM, CNN, and GNN models on EMBER 2018, EMBER 2024, LAMDA, and BenchMFC. Evaluation must cover three robustness protocols—cross-dataset, temporal, and regional—and report ROC-AUC, F1, MCC, robustness drop, and calibration error. Clean, well-commented code and structured experiment logs are essential because I am most concerned about the SLR’s depth and the correctness of the ML implementation. Deliverables • PRISMA-based SLR manuscript (Word or LaTeX) with all figures and appendices • Documented Python project (requirements.txt, notebooks or scripts, README) • CSV results, plots, and a brief slide deck summarising findings Timeline: preliminary draft and initial results by mid-October, final hand-over no later than 1 November. Budget is negotiable. When you reply, highlight your experience on similar research or malware-focused ML work; that is the main thing I will consider.