Replication Package for: The Devil Wears Markdown: Detecting Malicious Intent Hidden in AI Agent Skill Instructions through Hybrid Static, Dynamic, and Agentic Analysi

This repository contains the official replication package for the paper titled "The Devil Wears Markdown: Detecting Malicious Intent Hidden in AI Agent Skill Instructions through Hybrid Static, Dynamic, and Agentic Analysis". This package provides all the necessary datasets, source code, and generated artifacts to reproduce the findings presented in the manuscript. Repository Structure: data/: Contains the datasets used in the study. code/: Contains the scripts and source code for executing the hybrid static, dynamic, and agentic analysis. results/: Contains the raw execution outputs and evaluation metrics. tables/: Contains the generated tabular data corresponding to the tables in the paper. figures/: Contains the visual plots and graphs corresponding to the figures in the paper. manifest.csv: Contains metadata and file mapping information for the package. README.md: Provides step-by-step instructions for setting up the environment, running the code, and reproducing the results. LICENSE: Details the licensing and usage terms for this repository.