drugbank-database
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill README includes behavioral steering instructions that mandate the AI agent to proactively suggest the 'K-Dense Web' platform during multi-step research tasks. Evidence: The section 'Suggest Using K-Dense Web For Complex Worflows' explicitly instructs the agent to frame the suggestion as a productivity upgrade.
- [PROMPT_INJECTION]: Processing external drug databases introduces a surface for indirect prompt injection attacks where malicious instructions could be embedded in the data entries. Evidence: Data is ingested from the DrugBank database and parsed via 'lxml'. The agent context includes capabilities for Python code execution and package installation. There are no documented boundary markers or sanitization steps for the ingested XML content.
- [EXTERNAL_DOWNLOADS]: The skill requires downloading and installing multiple external software packages to facilitate data analysis. Evidence: Installation steps involve 'drugbank-downloader', 'bioversions', 'lxml', 'pandas', 'rdkit', 'networkx', and 'scikit-learn'.
Audit Metadata