bio-pathway-kegg-pathways
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill processes external data from 'de_results.csv'. While no sanitization steps are explicitly shown, the data is processed by specialized bioinformatic functions like 'enrichKEGG' which treat input as data rather than instructions.
- [Data Exposure & Exfiltration] (SAFE): The skill requires internet access to query the KEGG database and may open the KEGG website in a browser via 'browseKEGG'. These are documented, primary functions of the skill and do not target sensitive system files.
- [Unverifiable Dependencies] (SAFE): The skill relies on well-known Bioconductor R libraries ('clusterProfiler', 'org.Hs.eg.db', 'pathview') which are standard in the bioinformatics domain.
Audit Metadata