twilio-sendgrid-email-send

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill includes strong safety guidelines instructing the agent to always confirm recipients, subject, and content with the user before sending. This is a recommended safety pattern for communication-based skills.
  • [DATA_EXFILTRATION]: No evidence of unauthorized data access or transmission. The skill uses standard SendGrid API endpoints (api.sendgrid.com). It correctly recommends using os.environ["SENDGRID_API_KEY"] for secret management rather than hardcoding credentials.
  • [COMMAND_EXECUTION]: The provided Python and Node.js code snippets perform legitimate API calls using official libraries (sendgrid, @sendgrid/mail) and the requests library. No arbitrary command execution or shell injection risks were identified.
  • [EXTERNAL_DOWNLOADS]: The skill uses standard package managers and official SendGrid libraries. No suspicious remote code downloads or execution patterns were detected.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes dynamic template data and email content, it includes explicit instructions for the agent to verify content with the user before execution, mitigating the risk of instructions embedded in data being executed without oversight.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 07:12 AM
Security Audit — agent-trust-hub — twilio-sendgrid-email-send