openevidence-install-auth
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to install the official
@openevidence/sdkNode.js package and refers to theopenevidencePython package. These are standard operations for integrating the service. - [INDIRECT_PROMPT_INJECTION]: The skill sets up a client to process natural language medical queries (e.g., "What are first-line treatments for Type 2 diabetes?"). This constitutes a data ingestion surface for the agent. However, as this is the core intended functionality of a medical decision support tool and includes no execution capabilities on the ingested data, the risk is negligible.
- [CREDENTIALS_UNSAFE]: The skill correctly instructs users to use environment variables (
OPENEVIDENCE_API_KEY) and.envfiles for secret management rather than hardcoding credentials, which is a recommended security practice.
Audit Metadata