openevidence-multi-env-setup
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process data from external API responses (api.openevidence.com) and local input files (synthetic-query.json), creating a vulnerability surface where maliciously crafted external content could influence agent behavior.
- Ingestion points: Data received from openevidence.com API endpoints and contents of local JSON query files referenced in deployment workflows.
- Boundary markers: The provided code templates and instructions do not implement explicit delimiters or instructions to ignore embedded commands within the ingested data.
- Capability inventory: The skill utilizes curl for network interaction and npm run deploy for environment deployment, and is granted tools for Read, Write, Edit, and Grep operations.
- Sanitization: Verification logic focuses on ensuring required environment variables are set but does not include sanitization or validation of the content returned from external API calls.
- [COMMAND_EXECUTION]: The skill documentation includes instructions for executing shell commands, specifically curl for performing clinical queries and audit log verification, and npm for deployment processes.
- [DATA_EXFILTRATION]: The skill performs network operations to openevidence.com domains (including api.dev.openevidence.com and api.staging.openevidence.com) to manage environment configurations. While these domains are not on the standard whitelist, the operations are consistent with the skill's stated purpose of facilitating OpenEvidence platform integration.
Audit Metadata