The Agent Skills Directory

Remote Script Execution: The skill references a common installation pattern for Nextflow using curl -s https://get.nextflow.io | bash. While this is the standard and official method provided by the software vendor to set up the environment, piping remote content directly to a shell is a security consideration that should be reviewed.
Indirect Prompt Injection Surface: The skill ingests metadata from external bioinformatics databases (NCBI GEO/SRA and EBI ENA) to identify and download sequencing data.
Ingestion points: scripts/sra_geo_fetch.py and scripts/utils/ncbi_utils.py fetch data from NCBI's Entrez E-utilities and EBI's API.
Boundary markers: External data is processed via structured XML/JSON parsers rather than direct prompt interpolation, reducing the risk of accidental obedience to embedded instructions.
Capability inventory: The available capabilities include file system access, network operations, and command execution via subprocess and Nextflow.
Sanitization: Scripts implement specific regular expressions and field-matching logic to extract data from remote API responses.
System Command Execution: Scripts such as check_environment.py and manage_genomes.py utilize the subprocess module to interact with system tools like Docker, Java, and the AWS CLI. These calls are used for diagnostic purposes and managing genomic reference data.
External Repository Integration: The skill interacts with recognized public data repositories including NCBI, EBI, and public S3 buckets to download research data. These interactions are standard for the bioinformatics workflows the skill supports.

nextflow-development