primekg
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation provides instructions to download the PrimeKG dataset from Harvard Dataverse (dataverse.harvard.edu), which is a well-known and reputable service for scientific data hosting. It includes specific file IDs and expected MD5 checksums to ensure data integrity.- [COMMAND_EXECUTION]: The documentation includes shell command examples using 'curl' to facilitate the download of the required datasets. These are provided as user guidance and are not executed automatically by the skill's scripts.- [INDIRECT_PROMPT_INJECTION]: The skill processes external CSV data which is subsequently provided to the AI agent.
- Ingestion points: Data is loaded from a local CSV file in 'scripts/query_primekg.py' using 'pandas.read_csv'.
- Boundary markers: The data is returned to the agent as structured Python objects (dictionaries and lists), which limits the risk of the model interpreting data as instructions.
- Capability inventory: The skill's capabilities are restricted to local file reading and basic data manipulation; it has no network-write or arbitrary system command execution permissions.
- Sanitization: The script performs schema validation, verifying required columns and checking for empty values, and uses strict 'pandas' data types to ensure consistent parsing.- [SAFE]: No malicious patterns, obfuscation, or safety guideline bypass attempts were detected. The skill follows established practices for scientific data handling and local environment configuration.
Audit Metadata