The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill guides the user to install the @membranehq/cli package via npm to interface with the Membrane platform services.
[COMMAND_EXECUTION]: Employs shell commands using the membrane CLI to perform authentication, manage connection states, and execute email operations like reading, searching, and sending messages.
[REMOTE_CODE_EXECUTION]: Features the membrane action create command, which allows the agent to dynamically generate and deploy new integration logic on the remote Membrane platform based on natural language descriptions.
[PROMPT_INJECTION]: The skill processes external data from Gmail (messages and threads), which introduces a potential surface for indirect prompt injection if malicious instructions are embedded in ingested email content.
Ingestion points: Retrieves external email content using get-message, get-thread, and list-messages actions.
Boundary markers: No specific delimiters are defined in the instructions to isolate email data from the agent's primary instruction set.
Capability inventory: The skill possesses capabilities for data modification and deletion (e.g., delete-thread, send-message) which could be targeted via injection.
Sanitization: There is no mention of sanitization or filtering applied to the email body before it enters the agent's context.

gmail