email-compliance
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data such as email templates and signup forms (ingestion points). While it does not define specific prompt boundary markers, it mitigates risk by requiring verification via source code analysis and isolated staging environments. Capabilities are limited to observation and local sandboxing, and sanitization is enforced through a strict prohibition on sending emails to real addresses.
- [EXTERNAL_DOWNLOADS]: The skill provides links to official regulatory and standards documentation from well-known domains, including ftc.gov, w3.org, and various national government portals. These references are for manual research and do not involve automated script execution or remote code execution.
Audit Metadata