shopping
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources such as product pages, Instagram links, and Facebook Marketplace listings to evaluate products. \n
- Ingestion points: Data enters the context via browser search results, browser page verification, and social shopping links via
browserandclitools. \n - Boundary markers: The instructions do not specify explicit delimiters to isolate untrusted external content when it is processed by the agent. \n
- Capability inventory: The skill has high-impact capabilities including the ability to initiate and complete purchases through
shopify-ucp-cliand navigate the web viabrowser.spawn_task. \n - Sanitization: The workflow requires explicit user approval of a final quote before a purchase is finalized, which serves as a security checkpoint. \n- [COMMAND_EXECUTION]: The skill uses multiple platform-provided CLI tools (e.g.,
meta-catalog-search,facebook-cli,shopify-ucp-cli) to perform search, item retrieval, and checkout operations, incorporating user-provided strings as search queries.
Audit Metadata