vdm-next-audience-integration
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The instructions are focused on software engineering tasks and contain no patterns intended to bypass agent constraints or safety filters.
- [DATA_EXFILTRATION]: While the skill processes user identifiers such as emails and phone numbers, it explicitly defines a pipeline for normalization and cryptographic hashing (SHA256/SHA512) before data is sent to external APIs.
- [COMMAND_EXECUTION]: There are no instances of dynamic command execution, shell spawning, or subprocess calls involving untrusted data.
- [SAFE]: The skill uses industry-standard libraries like Zod for runtime type checking and validation, ensuring that only correctly structured data enters the transformation pipeline.
Audit Metadata