tooluniverse-precision-oncology
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill is designed to ingest large amounts of data from external sources, including clinical databases and research literature.
- Ingestion points: Data is retrieved from
PubMed_search_articles,BioRxiv_list_recent_preprints,MedRxiv_get_preprint, and various genomic APIs such as CIViC, COSMIC, and GDC as documented inSKILL.mdandTOOLS_REFERENCE.md. - Boundary markers: While the skill advises the agent to flag preprints as not peer-reviewed, it lacks explicit instructions to treat the content of external tool outputs as untrusted strings to prevent secondary injection attacks.
- Capability inventory: The skill possesses extensive data retrieval capabilities across dozens of specialized bioinformatics tools in its workflow. It does not have access to shell execution or direct file system manipulation beyond generating the output report file.
- Sanitization: The instructions do not specify any sanitization or filtering of external text content before it is incorporated into the synthesized clinical report.
Audit Metadata