a-b-test-config-creator
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No instructions designed to override AI safety constraints or bypass system rules were detected.
- [Data Exfiltration] (SAFE): No hardcoded secrets, API keys, or logic for accessing and exfiltrating sensitive local files were found.
- [Remote Code Execution] (SAFE): The skill contains no download commands (curl, wget) or patterns for executing remote scripts.
- [Command Execution] (SAFE): Although the skill lists 'Bash' as an allowed tool, it does not include any pre-configured malicious command strings.
- [Indirect Prompt Injection] (SAFE): The skill activates on specific strings but does not ingest and process untrusted external data in a way that creates a vulnerability surface based on the provided file.
Audit Metadata