The Agent Skills Directory

[SAFE]: The skill follows secure credential management practices by using environment variables (SENDGRID_API_KEY) for authentication.
[DATA_EXFILTRATION]: The skill transmits data to api.sendgrid.com, which is the official and documented API endpoint for the SendGrid service owned by the skill's author, Twilio.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data for email bodies and subjects.
Ingestion points: External data enters through the html_content, subject, and dynamic_template_data fields in SKILL.md.
Boundary markers: The instructions explicitly require the agent to confirm all email details with the user before sending, serving as a human-in-the-loop boundary.
Capability inventory: The skill utilizes the sendgrid and requests libraries for network communication and performs file reads for attachments.
Sanitization: Content sanitization is not specified in the code snippets; however, the skill mandates manual user review to mitigate risks of automated exploitation.

twilio-sendgrid-email-send