sorting-groceries
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data provided by the user, which presents a surface for indirect prompt injection.
- Ingestion points: User-provided images of aisle signs and text-based grocery lists are parsed and normalized in Step 1 and Step 3 of the workflow in SKILL.md.
- Boundary markers: The skill lacks explicit instructions for the model to ignore or treat as data any natural language instructions that might be embedded within the images or lists.
- Capability inventory: The skill's capabilities are restricted to text processing and output formatting within the agent's context. It lacks tools for network access, file system modification, or command execution.
- Sanitization: No validation or sanitization of the input text extracted from images is defined.
Audit Metadata