b2c-querying-data
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-controlled data such as search keywords, email addresses, and category IDs to construct database and search index queries.
- Ingestion points: Untrusted data enters the agent context through search phrases (
psm.setSearchPhrase), email filters inOrderMgr.searchOrdersandCustomerMgr.searchProfiles, and custom attribute filters inSystemObjectMgr.querySystemObjects. - Boundary markers: No explicit natural language boundary markers are defined in the prompt instructions.
- Capability inventory: The skill uses platform-specific data retrieval APIs including
ProductSearchModel.search(),OrderMgr.searchOrders(),CustomerMgr.searchProfiles(), andSystemObjectMgr.querySystemObjects(). - Sanitization: The provided code samples utilize the platform's native parameterized query syntax (e.g.,
'email = {0}', 'customer@example.com'), which effectively prevents query injection by separating the query logic from the data values. - [DATA_EXPOSURE]: The skill facilitates access to sensitive business objects, including Customer Profiles and Orders, which contain personally identifiable information (PII). The instructions focus on legitimate business logic for storefront and job processing and do not include any patterns for data exfiltration or unauthorized network transmission.
Audit Metadata