The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to call Poppler utilities (pdftotext and pdfinfo) for extracting content and metadata from PDF files. These calls are implemented securely using argument lists rather than shell strings, which mitigates the risk of command injection.
[EXTERNAL_DOWNLOADS]: The skill communicates with various well-known academic and machine learning platforms to fetch research data and datasets. This includes services such as arXiv, Semantic Scholar, PubMed, HuggingFace, Kaggle, and OpenML. These operations are necessary for the skill's stated purpose and target established repositories.
[PROMPT_INJECTION]: The skill processes text extracted from external PDFs and web pages. This presents an indirect prompt injection surface where malicious content within the processed documents could attempt to influence the agent's reasoning. However, this is an inherent characteristic of document analysis tools and does not indicate malicious intent in the skill's own code.

research