tooluniverse-gwas-finemapping
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security threats were detected in the provided documentation. The skill primarily acts as an educational guide and a conceptual wrapper for interacting with the Open Targets Genetics and GWAS Catalog APIs.\n- [PROMPT_INJECTION]: While the skill is designed to ingest data from external bioinformatics APIs (Open Targets and GWAS Catalog), which constitutes an indirect prompt injection surface, the risk is negligible as the skill lacks high-privilege capabilities such as arbitrary code execution, file system modification, or network exfiltration of sensitive local data.\n
- Ingestion points: External data is retrieved via GraphQL and REST API calls to genomics databases as described in the 'Tools Used' section of
SKILL.md.\n - Boundary markers: Not explicitly defined in the documentation snippets.\n
- Capability inventory: Functionality is limited to data retrieval, analysis of genetic variants, and generation of validation suggestions. No dangerous subprocess or shell execution capabilities are present in the documentation.\n
- Sanitization: Standard data handling for genomic identifiers (e.g., rsIDs, EFO IDs) is implied by the technical nature of the task.
Audit Metadata