shopify-admin-customer-win-back
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the Shopify API (customer names and emails) and includes it in agent-generated summaries and CSV files.
- Ingestion points: Customer names, emails, and order history from the
customersGraphQL query. - Boundary markers: The workflow lacks explicit delimiters or specific instructions for the agent to treat customer-provided strings strictly as data rather than instructions.
- Capability inventory: The skill can write to the local file system (CSV generation) and execute mutations (tagging) on the Shopify store via the shopify-admin-execution toolkit.
- Sanitization: No evidence of sanitization, filtering, or validation of the customer data is present before it is processed or stored.
Audit Metadata