The Agent Skills Directory

[SAFE]: No malicious prompt injection patterns or attempts to bypass safety filters were detected in the skill instructions or metadata. All instructions are focused on genetic research workflows.
[SAFE]: The skill does not contain hardcoded credentials, sensitive file access, or unauthorized network operations. External data is retrieved via specific tool calls to well-known scientific platforms (Open Targets Genetics and GWAS Catalog).
[SAFE]: The skill processes data from external genomic APIs in python_implementation.py, which represents a standard indirect prompt injection surface for biological data tools. This activity is essential to the skill's primary purpose and involves restricted, read-only capabilities.
Ingestion points: tu.run_one_function calls in _process_rsid, _process_gene, get_credible_sets_for_study, and search_gwas_studies_for_disease.
Boundary markers: Absent.
Capability inventory: Read-only API queries provided by the platform (ToolUniverse).
Sanitization: No explicit content filtering on biological data strings, as they are used for functional annotation ranking.
[SAFE]: No obfuscation, persistence mechanisms, or privilege escalation techniques were identified in the source code or documentation.

tooluniverse-gwas-finemapping