tooluniverse-cancer-variant-interpretation
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process large amounts of data from external sources, which presents a surface for indirect prompt injection.
- Ingestion points: The agent fetches data from external databases including PubMed (abstracts), CIViC (evidence items), OpenTargets (drug descriptions), and clinical trial registries in Phases 2, 4, 5, 6, and 7.
- Boundary markers: The instructions do not specify the use of boundary markers or instructions to the agent to disregard potential commands embedded within the fetched clinical literature or database descriptions.
- Capability inventory: The skill possesses file-writing capabilities (creating a markdown report in Phase 8) and extensive tool-calling permissions across multiple biological and medical domains.
- Sanitization: There is no mention of sanitizing or validating external content before it is interpolated into the final clinical report.
Audit Metadata