The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or safety filter bypass attempts were identified. The skill performs legitimate bioinformatics tasks using established libraries.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes data from external bioinformatics APIs (UniProt, KEGG, ChEMBL, etc.). While this provides a theoretical ingestion surface for indirect prompt injection, the sources are highly trusted public scientific institutions.
Ingestion points: External API data is fetched and processed in scripts/protein_analysis_workflow.py, scripts/compound_cross_reference.py, and scripts/pathway_analysis.py.
Boundary markers: No specific prompt-level delimiters are used for API data interpolation.
Capability inventory: The skill uses the Bash tool and performs local file writing for data exports.
Sanitization: Standard parsing is applied to scientific data formats (FASTA, KGML, TSV).
[EXTERNAL_DOWNLOADS]: The skill establishes network connections to well-known scientific databases to fetch biological information, which is the primary and documented purpose of the tool.
[COMMAND_EXECUTION]: Included Python scripts perform expected local operations such as directory creation (os.makedirs) and saving analysis results to CSV and SIF formats.

bioservices