phoenix-evals
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends installing official vendor packages (such as
arize-phoenixand@arizeai/phoenix-client) and well-known AI SDKs from providers like OpenAI, Anthropic, and Google. - [COMMAND_EXECUTION]: Provides code snippets for creating deterministic evaluators and running batch evaluations using the Phoenix Python and TypeScript libraries. All execution patterns are standard for the tool's intended purpose of AI performance monitoring.
- [SAFE]: No security issues detected. The skill does not contain hardcoded credentials, malicious obfuscation, or attempts to bypass LLM safety guidelines. It promotes security best practices, such as using XML delimiters in prompt templates to prevent indirect prompt injection.
Audit Metadata