The Agent Skills Directory

[SAFE]: The skill implements a standard, well-documented bioinformatics workflow for two-group differential expression analysis on bulk RNA-seq data. It includes comprehensive error handling and input validation.
[COMMAND_EXECUTION]: The scripts are designed to be run via Rscript. Analysis of the code shows no unsafe shell execution or dynamic command building using unvalidated user input. It uses standard R package loading mechanisms.
[DATA_EXFILTRATION]: No network operations or suspicious external data transmissions were identified. The code reads local CSV files and writes output (tables, plots, and session info) to a user-defined local directory.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied CSV files for expression data and group annotations. However, these are handled as structured data for statistical analysis. There are no mechanisms detected where data content is interpolated into agent instructions or executed as code, mitigating the risk of injection.
[OBFUSCATION]: The code is written in clear, maintainable R. No evidence of Base64 encoding, zero-width characters, or homoglyph-based obfuscation was found in the scripts or documentation.

deg-screening-analysis