The Agent Skills Directory

[SAFE]: The skill performs its stated function of scientific data analysis without any detected malicious behavior.- [DATA_EXPOSURE]: The skill reads user-provided scientific data files to extract metadata and perform statistical analysis. This access is restricted to the files provided for analysis and is necessary for the skill's primary purpose.- [REMOTE_CODE_EXECUTION]: The provided Python script (scripts/eda_analyzer.py) uses standard scientific libraries (e.g., NumPy, Pandas, Biopython) to parse data. While these libraries process external data, the script does not use insecure functions like eval() or exec(). The reference documentation mentions Python's pickle module for specific formats but correctly includes a warning about only using it with trusted sources.- [PROMPT_INJECTION]: No direct prompt injection or instructions to bypass safety guidelines were found in the skill metadata or instructions.- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and summarizes untrusted scientific data. However, the risk is minimized as it primarily performs numerical and structural analysis, and the generated reports are intended for human review. No signs of active exploitation or malicious redirection were identified.

exploratory-data-analysis