ab-test-setup
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill file
SKILL.mdcontains only structured markdown text providing guidance on statistical rigor and process flow for A/B testing. There are no scripts, command-line instructions, or platform-specific dynamic contexts. - [INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting user input regarding test hypotheses and metrics. While this creates a data ingestion surface, the skill lacks any executable capabilities (such as network access, file writing, or shell execution) that could be leveraged for an attack.
- Ingestion points: Sections 3, 4, 6, and 7 involve processing user-provided test parameters.
- Boundary markers: Includes logical gates and explicit confirmation prompts (e.g., section 3) to ensure user intent matches the processed data.
- Capability inventory: No capabilities detected across the skill.
- Sanitization: Not applicable as the data is used for advisory purposes only without execution.
Audit Metadata