bio-de-deseq2-basics
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- EXTERNAL_DOWNLOADS (MEDIUM): The skill executes
install.packages('BiocManager')andBiocManager::install()to downloadDESeq2andapeglm. These R repositories (CRAN and Bioconductor) are external and not included in the 'Trusted External Sources' whitelist, making these unpinned downloads unverified.- REMOTE_CODE_EXECUTION (MEDIUM): Runtime installation and loading of R packages allows for the execution of code from external sources. The absence of version locking or hash verification increases the risk of supply chain attacks.- DATA_EXFILTRATION (LOW): The skill useswrite.csvto save results to the local filesystem. While this is standard for the tool's purpose, it represents a capability to write data that could be redirected to sensitive paths if the filenames were manipulated.- INDIRECT_PROMPT_INJECTION (LOW): The skill processes external data that could serve as an attack surface. - Ingestion points: Untrusted data enters via the
countDatamatrix,colDatametadata frame, and external files loaded viatximport. - Boundary markers: None. The data is processed directly as R objects.
- Capability inventory: The skill has file-writing capabilities (
write.csv). - Sanitization: There is no evidence of sanitization for metadata strings (e.g., column names in
coldata) which are often interpolated into design formulas.
Audit Metadata