email-with-resend

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill describes a system for processing external data through email templates and webhooks, which represents a potential injection surface.
  • Ingestion points: Untrusted data enters the context via template variables (e.g., user names, URLs) and incoming webhook payloads from Resend.
  • Boundary markers: The instructions explicitly mandate the use of React Email or provider-side templates to ensure structural separation and include instructions to escape every untrusted value.
  • Capability inventory: The skill does not perform any direct file writes, network requests, or command executions; it provides guidance for implementation within the user's codebase.
  • Sanitization: There is a strong emphasis on sanitizing external content, including manual escaping of HTML and text rendering.
  • [CREDENTIALS_SAFE]: The skill demonstrates best practices for secret management.
  • The implementation guide uses process.env.RESEND_API_KEY and process.env.RESEND_WEBHOOK_SECRET for credential access, avoiding hardcoded secrets or unsafe storage instructions.
  • [DATA_EXPOSURE]: The skill provides explicit privacy-preserving guidance.
  • It warns against exposing raw email addresses in URLs (e.g., for unsubscribe links), recommending the use of opaque, signed identifiers or tokens to prevent PII leakage.
  • [COMMAND_EXECUTION]: All provided code snippets are standard application-level logic for a Node.js/TypeScript environment.
  • There are no instances of eval(), exec(), or other dangerous dynamic execution patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 06:57 AM
Security Audit — agent-trust-hub — email-with-resend