alterlab-vcd-packaging
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's architecture involves ingesting data from external sources, which constitutes a vulnerability surface for indirect prompt injection. \n
- Ingestion points: The agent is explicitly instructed to search the web for packaging trends, competitor designs, and regulatory requirements (Step 1), and to read existing project files such as brand guidelines and product specifications (Step 1 and Step 4). \n
- Boundary markers: The skill body lacks any boundary markers, delimiters, or explicit instructions to the agent to disregard or treat with suspicion any natural language instructions that might be embedded in the search results or files it processes. \n
- Capability inventory: The skill has the capability to read and write several markdown files to the project directory, including structural specifications, print production specifications, and unboxing experience maps. \n
- Sanitization: There are no documented mechanisms for sanitizing, filtering, or validating the external content retrieved during the research phase before it is used to influence the agent's design decisions and file generation.
Audit Metadata