skills/openclaw/skills/send-email/Gen Agent Trust Hub

send-email

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides legitimate instructions and production-ready code examples for integrating with the Resend email service across multiple programming languages including Node.js, Python, and Go.
  • [EXTERNAL_DOWNLOADS]: Installation instructions utilize standard, well-known package managers (npm, pip, cargo) to fetch official SDKs from established registries.
  • [CREDENTIALS_UNSAFE]: The documentation follows security best practices by recommending the use of environment variables (RESEND_API_KEY) and webhook secrets (RESEND_WEBHOOK_SECRET) rather than hardcoding sensitive credentials in source code.
  • [PROMPT_INJECTION]: The skill processes external data via webhooks, which constitutes a potential surface for indirect prompt injection if the agent takes subsequent actions based on unvalidated event content. 1. Ingestion points: references/webhooks.md (Webhook event payloads from Resend). 2. Boundary markers: Cryptographic signatures (svix-signature) are used to verify authenticity, but no specific natural language delimiters or 'ignore instructions' markers are mentioned for the payload content. 3. Capability inventory: references/webhooks.md (Updating databases and triggering follow-up application logic based on delivery events). 4. Sanitization: Mandatory signature verification is documented to ensure data integrity and prevent spoofing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 02:02 PM