scenic-gene-regulatory-network
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or credential safety issues were detected within the skill files.
- [EXTERNAL_DOWNLOADS]: Fetches species-specific ranking databases and motif annotations from the Aerts Lab (University of Leuven) resource server (resources.aertslab.org). These downloads are legitimate requirements for the stated functionality of the Gene Regulatory Network pipeline.
- [COMMAND_EXECUTION]: Provides routine environment setup commands for directory creation (
mkdir) and data retrieval (wget) using standard system utilities. - [PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection as it processes user-provided expression data and external databases.
- Ingestion points: User-provided expression data (
adata) and external reference files (.feather,.tbl). - Boundary markers: The instructions specify strict data format requirements (raw counts) and species-specific gene naming conventions.
- Capability inventory: Includes file system operations (
mkdir,write,to_csv,pickle.dump), network retrieval (wget), and parallel processing vian_jobs. - Sanitization: Includes a dedicated 'Defensive Validation Patterns' section providing Python logic to verify data scales, file existence, and gene nomenclature before processing.
Audit Metadata