The Agent Skills Directory

[Command Execution] (SAFE): The skill documents the use of pysam.samtools and pysam.bcftools. These functions execute external bioinformatics utilities as part of their intended primary purpose. The examples provided use static file paths and do not demonstrate unsafe interpolation of untrusted data.
[External Downloads] (SAFE): The skill includes instructions to install the pysam package using standard package managers (uv pip install pysam). pysam is a reputable and standard library in the scientific community.
[Data Exposure] (SAFE): The skill interacts with genomic data formats (BAM, VCF, FASTA) and standard input/output streams. No patterns suggesting the exfiltration of sensitive system files or credentials were found.
[Indirect Prompt Injection] (SAFE): While the skill ingests external data files, these are highly structured genomic formats. The risk of these files containing adversarial instructions that could manipulate the agent is negligible.

pysam