The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the membrane CLI to perform operations such as searching for and executing actions. This is the intended functionality of the skill and uses a controlled interface for API interactions.
[EXTERNAL_DOWNLOADS]: The instructions guide the user to install the @membranehq/cli package via NPM. This is a standard practice for utilizing the Membrane platform and the package is a recognized vendor resource.
[CREDENTIALS_UNSAFE]: The skill explicitly advises against asking users for API keys or tokens, instead directing them to use a secure connection-based authentication flow managed by the platform.
[DATA_EXFILTRATION]: No unauthorized data transmission patterns were detected. Network access is scoped to the official Membrane and Cascade Strategy endpoints.

cascade-strategy