The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires installing the @membranehq/cli package from the NPM registry and also uses npx to run commands. These are official tools from the skill author's organization.
[COMMAND_EXECUTION]: Instructions include several shell commands for logging in, ensuring connections, and running actions. The membrane request feature allows the agent to execute arbitrary HTTP requests against the Google Drive API.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from Google Drive which could contain malicious instructions. 1. Ingestion points: Metadata, file lists, and comments retrieved via the Drive API. 2. Boundary markers: No explicit delimiters or instructions are provided to the agent to treat Drive content as untrusted. 3. Capability inventory: Capabilities include deleting files, modifying permissions, and performing raw API requests. 4. Sanitization: The skill does not specify any sanitization or validation logic for the external data it processes.

google-drive