The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation includes standard installation instructions for the 'polars-bio' and 'polars' Python packages via official package managers (pip, uv). These are legitimate, purpose-built libraries for the functionality described.
[COMMAND_EXECUTION]: The skill provides various commands for reading, writing, and processing genomic data files locally and on cloud storage. All operations are within the scope of a bioinformatics tool and follow industry-standard practices for handling large-scale data.
[DATA_EXFILTRATION]: While the skill interacts with cloud storage services (Amazon S3, Google Cloud Storage, Azure Blob Storage), these operations are transparently documented as part of the tool's core functionality for data access. No unauthorized or suspicious network operations to unknown domains were detected.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface through its various file readers (BED, VCF, BAM, etc.) and SQL interface. While this represents a theoretical surface for indirect prompt injection if an agent is instructed to follow commands embedded in data, it is a structural characteristic of data processing tools and no malicious use-case is suggested.
Ingestion points: Multiple functions in 'references/file_io.md' and 'references/sql_processing.md' facilitate reading from external files and cloud endpoints.
Boundary markers: Not explicitly defined within the documentation for processed data content.
Capability inventory: File system access and cloud network operations for data retrieval/storage.
Sanitization: Standard data validation for file formats is implied, but semantic sanitization of processed strings is not specified.

polars-bio