ab-testing
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill provides instructional content and templates for A/B testing. It contains no scripts or code, relying entirely on markdown instructions for the AI agent.
- [INDIRECT_PROMPT_INJECTION]: The instructions direct the agent to read local context files (e.g.,
.agents/product-marketing.md) to gain business context. This represents a data ingestion surface, but since the skill has no tools for file writing, network exfiltration, or code execution, the risk is negligible. - Ingestion points:
.agents/product-marketing.md,.claude/product-marketing.md, andproduct-marketing-context.mdmentioned in SKILL.md. - Boundary markers: None.
- Capability inventory: None; the skill is purely informational markdown.
- Sanitization: None.
- [EXTERNAL_DOWNLOADS]: The skill references several external A/B testing calculators and tools via links in the documentation.
- Evidence: Links to
evanmiller.org,optimizely.com,abtestguide.com, andvwo.comin SKILL.md and references/sample-size-guide.md. These are well-known industry services used for informational and calculation purposes.
Audit Metadata