The Agent Skills Directory

[SAFE]: No malicious patterns, exfiltration attempts, or obfuscation were detected. The skill provides architectural patterns for email systems, such as queuing, template versioning, and bounce handling logic.
[INDIRECT_PROMPT_INJECTION]: The skill documents ingestion of untrusted data (e.g., user profiles, tokens) into email templates.
Ingestion points: references/detailed-guide.md (template interpolation patterns).
Boundary markers: Not explicitly defined for template placeholders.
Capability inventory: Sends emails via external providers; no file-system writes or shell execution documented.
Sanitization: Not explicitly implemented in the provided pseudocode, but recommended as a general practice for email systems.
[EXTERNAL_DOWNLOADS]: The guide references well-known packages such as @react-email/components and html-to-text, which are standard for the described use case.

email-systems