policy-ai-content-ethics
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines an ingestion surface by interpolating user-provided inputs, such as business names and industries, into a pre-defined policy template using bracketed placeholders. While this is an attack surface for indirect prompt injection, the skill lacks tool-use capabilities, network access, or code execution, meaning any injection would be confined to the generated text and cannot lead to system compromise or data exfiltration.
- [SAFE]: The skill incorporates robust security instructions, explicitly prohibiting the input of Personally Identifiable Information (PII) or trade secrets into cloud AI prompts. It also includes an internal checklist for consultants to monitor and report jailbreak attempts, demonstrating a high degree of security awareness and protective guidance.
Audit Metadata