tooluniverse-drug-drug-interaction
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and analyze medication lists and patient safety data, which are external, untrusted inputs. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the drug names or medical descriptions to influence the agent's behavior or reporting.
- Ingestion points: Medication names, polypharmacy lists, and clinical data provided by the user in the prompt context.
- Boundary markers: The instructions do not define clear delimiters or specific instructions for the agent to ignore any natural language commands found within the medication data.
- Capability inventory: The skill involves writing findings to local files (DDI_risk_report.md) and implies the use of external search or database tools to extract clinical evidence.
- Sanitization: There is no evidence of input validation, schema enforcement, or sanitization of the drug identifiers before they are used in report generation or tool calls.
Audit Metadata