enhance-onboarding
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted external data by instructing the agent to analyze existing product UIs and onboarding flows to suggest improvements. While this represents a surface for indirect prompt injection, the skill's capabilities are limited to providing design recommendations and adding analytics instrumentation, which minimizes the threat.
- Ingestion points: Product landing pages, signup flows, and existing analytics funnels mentioned in the reasoning and procedure sections.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to disregard embedded content in the analyzed UIs.
- Capability inventory: The skill utilizes analytics SDK instrumentation and Playwright for new-user path verification.
- Sanitization: The skill mandates that no PII (Personally Identifiable Information) be included in analytics properties and that no events fire before consent is obtained.
- [SAFE]: The skill is primarily instructional and promotes standard UX/UI best practices. It includes a clear worked example, regression testing requirements, and safety guardrails that emphasize data privacy and user consent.
Audit Metadata