alterlab-vcd-packaging
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No security issues detected. The skill instructions follow best practices for a packaging design specialist, focusing on structural dielines and print production details. It does not request elevated privileges or execute code.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it is designed to ingest and process external information, although the overall risk is minimal.
- Ingestion points: The skill reads project-related files such as brand guidelines, product specifications, and regulatory documentation, and it searches the web for market trends.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to distinguish between its core instructions and data processed from external sources.
- Capability inventory: The agent's capabilities are limited to web searching and writing structured markdown files to the project directory.
- Sanitization: No sanitization or validation mechanisms are specified for content retrieved from external files or web searches.
Audit Metadata