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
Risk Level
SAFE
Analyzed
Jun 16, 2026, 10:27 AM
Security Audit — agent-trust-hub — llm-as-judge