llm-as-judge
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides educational content focused on evaluation methodology. The prompt templates use XML-style boundary markers (e.g., <user_query>, ) to separate instructions from untrusted content, effectively following best practices for preventing indirect prompt injection.
- [SAFE]: No patterns of data exfiltration, credential harvesting, or obfuscation were detected. Code snippets are limited to standard LLM API interactions and local metric calculations using established libraries like scikit-learn.
- [SAFE]: The skill does not perform unauthorized network operations, file system modifications, or command execution.
Audit Metadata