tooluniverse-rare-disease-diagnosis
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns detected. The skill maintains a clear focus on its stated purpose of clinical research and diagnosis support without evidence of obfuscation or unauthorized access.
- [EXTERNAL_DOWNLOADS]: The skill retrieves biological data and research findings from trusted organizations and well-known services, including PubMed, BioRxiv, ClinVar, and Orphanet.
- [PROMPT_INJECTION]: Evaluated for indirect prompt injection vulnerability surfaces in the processing of patient phenotypes and literature search results.
- Ingestion points: User-provided clinical symptoms in Phase 1 and metadata from research articles in Phase 6.
- Boundary markers: Clinical data is delimited by being converted into standardized HPO terms and structured into diagnostic report sections.
- Capability inventory: The skill's operations are confined to informational database queries and creating diagnostic reports on the local filesystem.
- Sanitization: While explicit sanitization is not detailed, the risk is inherently low as the skill interprets data within a highly specific clinical context.
Audit Metadata