skills/aipoch/medical-research-skills/conventional-oncology-hub-gene-research-planner/Gen Agent Trust Hub
conventional-oncology-hub-gene-research-planner
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied cancer types and research directions to formulate study designs, creating an ingestion point for external data.
- Ingestion points: User input strings identifying the cancer type and biomarker direction (SKILL.md).
- Boundary markers: Present. The skill defines specific input validation rules and a mandatory redirect for clinical or out-of-scope requests to prevent inappropriate medical usage.
- Capability inventory: None. The skill generates text-based planning documentation and does not perform executable code runs, file system modifications, or unauthorized network operations.
- Sanitization: The skill uses semantic scope checking to ensure user requests align with bulk-transcriptome study patterns before processing.
Audit Metadata