tooluniverse-variant-interpretation
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, prompt injections, or unauthorized data access were identified. The redundant YAML frontmatter appears to be a formatting error and contains no hidden payloads.\n- [DATA_EXFILTRATION]: The skill aggregates data from well-known scientific and clinical resources such as ClinVar, gnomAD, OMIM, ClinGen, and PubMed. These network operations are strictly related to the skill's stated purpose of variant interpretation and target established government and academic domains.\n- [PROMPT_INJECTION]: Instructions regarding 'English-first queries' are functional guidelines to ensure tool reliability and do not involve bypassing safety filters or overriding core agent behavior.\n- [REMOTE_CODE_EXECUTION]: The skill utilizes external scientific APIs and services like SpliceAI and Nvidia AlphaFold2. These interactions occur through defined tool interfaces and do not involve arbitrary code execution or downloading unverified executable content.
Audit Metadata