The Agent Skills Directory

[DATA_EXPOSURE]: The skill is designed to interact with a remote MCP server hosted by the vendor (Superhuman) to access email and calendar data. This is the primary functionality and is documented clearly as being sent to the AI client for reasoning.
[INDIRECT_PROMPT_INJECTION]: The skill handles untrusted external data in the form of email threads, message contents, and attachments. It provides effective mitigation by instructing the agent to summarize intended actions and obtain user confirmation before performing any write or destructive operations (e.g., sending mail, trashing threads).
[SAFE]: The skill follows security best practices by recommending the creation of drafts when intent is ambiguous and ensuring specific user authorization for reading sensitive data like inbox searches or contact details.

superhuman-mail