The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from genomic file formats (SAM, BAM, VCF, BCF, FASTA, FASTQ).
Ingestion points: Data enters the agent's context through pysam.AlignmentFile, pysam.VariantFile, pysam.FastaFile, pysam.FastxFile, and pysam.TabixFile as documented in SKILL.md and the reference files.
Boundary markers: There are no explicit boundary markers or safety instructions directing the agent to ignore potential instructions embedded within the genomic sequence data, read names, or variant metadata.
Capability inventory: The skill allows for file writing (AlignmentFile.write, VariantFile.write, and standard Python open) and the execution of external bioinformatics commands via pysam.samtools and pysam.bcftools wrappers.
Sanitization: No sanitization or validation logic is provided to filter out potential malicious instructions embedded in the genomic data.
[COMMAND_EXECUTION]: The skill provides instructions and examples for executing external command-line utilities like samtools and bcftools using the pysam library wrappers. These tools execute as subprocesses and can perform significant file system and data processing operations.

pysam