skills/aipoch/medical-research-skills/bioinformatics-translational-opportunity-finder/Gen Agent Trust Hub
bioinformatics-translational-opportunity-finder
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [SAFE]: The skill consists of markdown instructions and reference modules intended for scientific positioning. It does not ship with executable scripts or perform sensitive system operations. All referenced external materials trace back to the author's official GitHub repository.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied bioinformatics data (Ingestion Point: Step 1 in SKILL.md). While this is a surface for indirect prompt injection, the skill employs Step 1.5 check-ins and Section A boundaries as markers, has no dangerous capabilities (Inventory: None), and provides an explicit out-of-scope redirect for clinical misuse (Sanitization: "Out-of-scope" section in SKILL.md).
- [METADATA_POISONING]: An evaluation report (eval_report_bioinformatics-translational-opportunity-finder_result.json) is included, containing claims of its own security and stability. These self-referential safety claims were identified as metadata data points and were disregarded in favor of an independent audit, which confirmed the skill's safety.
Audit Metadata