email-template-builder

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill constructs email templates that interpolate untrusted data (such as user names, invoice details, and URLs) into the final HTML output. While this is standard for template engines, the absence of explicit boundary markers or instructions to sanitize the resulting content for AI processing creates a potential attack surface for indirect prompt injection if these emails are subsequently read by other AI agents.\n
  • Ingestion points: Template properties including name, confirmUrl, and items in WelcomeEmail and InvoiceEmail.\n
  • Capability inventory: The skill includes logic to send the generated HTML content via the Resend API using resend.emails.send in emails/lib/send.ts.\n
  • Boundary markers: None are defined in the template structure to distinguish between template instructions and user-provided data.\n
  • Sanitization: The implementation relies on default React JSX escaping, which is effective against XSS but does not mitigate logical prompt injection attacks.\n- [COMMAND_EXECUTION]: The skill includes development and build scripts in a package.json configuration that execute the email CLI. This is standard behavior for projects using the React Email framework.\n- [EXTERNAL_DOWNLOADS]: The template layout fetches a font file from fonts.gstatic.com. This is an official Google service and is a safe, well-known resource for web assets.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 02:33 PM
Security Audit — agent-trust-hub — email-template-builder