alterlab-primekg

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes Bash(python:*) and Bash(uv:*) to execute its data processing scripts, which is the intended method for interacting with the knowledge graph data.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes structured biomedical data from a local CSV file (kg.csv). While this constitutes a data ingestion surface, the risk is minimal as the script performs standard filtering and retrieval without executing data content.
  • Ingestion points: scripts/query_primekg.py reads the dataset using pd.read_csv.
  • Boundary markers: None.
  • Capability inventory: Python script execution via Bash.
  • Sanitization: Standard pandas filtering logic is applied to the data.
  • [EXTERNAL_DOWNLOADS]: The skill documentation references the PrimeKG dataset from Harvard Dataverse. The scripts themselves do not perform remote downloads, relying instead on a locally provided file path via the PRIMEKG_DATA_PATH environment variable.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 02:36 PM