The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes a wide variety of scientific data files including genomics, chemistry, and imaging formats. Evidence:
Ingestion points: File reading logic in scripts/eda_analyzer.py and direct agent interactions with file content.
Boundary markers: Not explicitly implemented in the provided scripts.
Capability inventory: Filesystem read and write access for generating markdown reports.
Sanitization: Relies on standard scientific parsing libraries (e.g., Pandas, Biopython).
[COMMAND_EXECUTION]: The skill includes a local analysis script scripts/eda_analyzer.py. This script performs structured data extraction and statistical analysis using established libraries and does not invoke arbitrary shell commands.
[EXTERNAL_DOWNLOADS]: The reference files mention various specialized scientific libraries such as pysam, rdkit, and tifffile. These are well-known, reputable packages in the scientific community intended for installation via official package managers like pip or conda. No automated or hidden remote code downloads or execution patterns were detected in the skill code.

exploratory-data-analysis