The Agent Skills Directory

[SAFE]: The skill implements functionality for a well-known service using official SDKs and standard API patterns. No malicious behavior, obfuscation, or persistence mechanisms were detected.
[PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection by allowing the agent to ingest and process inbound email content.
Ingestion points: Inbound email bodies are retrieved using resend.emails.receiving.get() as shown in references/receiving.md.
Boundary markers: While not implemented in the code snippets, the SKILL.md file provides a critical warning advising users to utilize the agent-email-inbox skill for secure processing of untrusted content.
Capability inventory: The skill has the capability to send emails via resend.emails.send, manage contact lists, and modify domain configurations.
Sanitization: Examples show the retrieval of raw HTML and text content without specific sanitization steps, relying on the user to handle the data safely.
[EXTERNAL_DOWNLOADS]: The skill includes code for downloading email attachments from the Resend API.
Evidence: references/receiving.md demonstrates fetching content from attachment.download_url.
Context: The downloads target the official Resend API infrastructure and are a necessary component for email receiving functionality.

resend