ab-testing
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill references several external resources, including A/B testing calculators and platforms (Evan Miller, Optimizely, VWO, and AB Test Guide). These are well-known services in the experimentation community and are used here for legitimate informational purposes.
- [PROMPT_INJECTION]: The skill instructions suggest reading product marketing context from local files (e.g.,
.agents/product-marketing.md). While this represents a surface for indirect prompt injection, the skill does not possess any dangerous capabilities (such as command execution or file-writing tools) that could be leveraged for an attack. The ingestion points are clearly defined and the risk is negligible. - [SAFE]: All tools mentioned (PostHog, LaunchDarkly, Split) are cited as options for the user to implement A/B testing in their own environment; the skill does not attempt to install or execute these tools automatically.
Audit Metadata