tooluniverse-variant-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions are aligned with its stated purpose of bioinformatics analysis. No evidence of malicious behavior, prompt injection, or unauthorized system access was found.
- [EXTERNAL_DOWNLOADS]: The skill references established and reputable bioinformatics services including ClinVar, dbSNP, gnomAD, Ensembl, and MyVariant.info for variant annotation. These well-known services are integral to the skill's function and are accessed through documented tool interfaces.
- [INDIRECT_PROMPT_INJECTION]: The skill processes VCF files, which are external data sources that may contain malicious text in metadata or info fields, representing an inherent surface for indirect prompt injection.
- Ingestion points: VCF files processed via the
parse_vcfandparse_vcf_cyvcf2functions (Phase 1). - Boundary markers: The skill documentation does not specify markers or delimiters to protect the agent's context from embedded instructions in variant data.
- Capability inventory: The skill utilizes tools to query external bioinformatics APIs and generates formatted markdown reports.
- Sanitization: There are no explicit instructions for sanitizing or filtering external file content before it is processed by the agent.
Audit Metadata