email-template-builder
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to ingest untrusted external data (such as names, invoice numbers, and item descriptions) to generate HTML email templates that are subsequently sent via the Resend API. This workflow creates an inherent surface for indirect prompt injection if the input data contains malicious instructions or payloads. Although the skill utilizes React, which provides default property escaping, the lack of explicit validation or boundary markers for interpolated data remains a minor risk factor.
- Ingestion points: Data properties passed to components in
emails/templates/welcome.tsxandemails/templates/invoice.tsx. - Boundary markers: Absent; the system relies on standard JSX interpolation without explicit delimiters for external content.
- Capability inventory: The skill possesses network communication capabilities through the Resend API in
emails/lib/send.ts. - Sanitization: The implementation benefits from React's built-in escaping of dynamic values to prevent basic HTML injection, though it does not implement secondary validation for instructions hidden within text strings.
Audit Metadata