agentmail-sending
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides example
curlcommands for interacting with the AgentMail API. These are standard implementation examples intended for the agent to follow and do not execute automatically on skill load. - [DYNAMIC_EXECUTION]: Includes a
script_upsertexample that demonstrates how to create a reusable email-sending script. This follows the standard pattern for workflow automation on the platform. - [DATA_EXFILTRATION]: The skill's primary purpose is to send data (email content) to the external service
api.agentmail.to. This is the documented and intended behavior, and the skill explicitly requires human approval and oversight for external communications. - [CREDENTIALS_UNSAFE]: Code examples correctly use environment variables and the platform's internal configuration system to manage the
AGENTMAIL_API_KEY, adhering to safe secret management practices.
Audit Metadata