The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses npx to execute shell commands via the @useclawlink/cli tool for account authentication, connection management, and performing Gmail actions.
[EXTERNAL_DOWNLOADS]: The instructions direct the agent to run npx @useclawlink/cli, which downloads and executes a package from the NPM registry at runtime.
[CREDENTIALS_UNSAFE]: The skill manages authentication by storing credentials in the local file system at ~/.clawlink/credentials.json. This is the standard storage location for this vendor's CLI tool.
[INDIRECT_PROMPT_INJECTION]: The skill provides tools to ingest untrusted data from an external source (Gmail inbox) which could contain malicious instructions.
Ingestion points: The skill reads external content through actions like gmail_fetch_emails, gmail_fetch_message_by_message_id, gmail_fetch_message_by_thread_id, gmail_list_drafts, gmail_get_attachment, and gmail_list_threads (SKILL.md).
Boundary markers: The skill documentation includes a note to "confirm with the user before any write," acting as a manual boundary for destructive actions.
Capability inventory: The skill possesses write capabilities including gmail_send_email, gmail_create_email_draft, gmail_reply_to_thread, gmail_forward_message, gmail_create_label, and gmail_add_label_to_email (SKILL.md).
Sanitization: There is no evidence of automated sanitization or filtering of the ingested email content before it is processed by the agent.

gmail