The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by reading Gmail messages, threads, and drafts. This content could contain malicious instructions designed to hijack the agent's logic.
Ingestion points: gmcli <email> search, gmcli <email> thread, and gmcli <email> drafts list in SKILL.md.
Boundary markers: None specified to delimit or neutralize untrusted email content.
Capability inventory: The skill can send emails, manage drafts, and execute shell commands.
Sanitization: No sanitization, filtering, or escaping of external email content is described.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of a third-party Node.js package (@mariozechner/gmcli) from the public npm registry.
[COMMAND_EXECUTION]: The skill operates by executing system shell commands to interact with the gmcli tool for searching, reading, and sending emails.

gmcli