skills/aipoch/medical-research-skills/single-gene-oncology-reference-grounded-research-planner/Gen Agent Trust Hub
single-gene-oncology-reference-grounded-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 untrusted user input (cancer types, genes, and research directions) to drive multi-step planning logic and literature retrieval. There are no explicit boundary markers or instructions to disregard potential injection attempts within the user-provided strings.
- Ingestion points: User input parsed in Step 1 of
SKILL.md(e.g., [cancer type] + [target gene]). - Boundary markers: Absent; user input is interpolated directly into the reasoning steps without delimiters.
- Capability inventory: Generates research designs, figure plans, validation hierarchies, and performs literature retrieval (Step 4.5).
- Sanitization: Absent; the skill relies on high-level scope checks but does not sanitize the content of input strings.
Audit Metadata