ab-test-setup
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No evidence of prompt injection patterns. The instructions focus entirely on providing a framework for A/B testing design.
- [DATA_EXFILTRATION]: No sensitive file access or unauthorized network operations detected. External links point to well-known industry resources for statistical calculations (Evan Miller and Optimizely).
- [REMOTE_CODE_EXECUTION]: No remote code execution or suspicious script downloads were identified. The skill is purely informational.
- [COMMAND_EXECUTION]: The skill does not contain any shell commands or requests for terminal execution.
- [CREDENTIALS_UNSAFE]: No hardcoded credentials, API keys, or secrets were found.
- [SAFE]: The skill provides structured guidance for experimentation, including hypothesis frameworks and metric selection, without introducing security risks.
Audit Metadata