The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the membrane command-line interface to manage connections and execute actions. This is the intended operational model for interacting with the Membrane ecosystem.
[EXTERNAL_DOWNLOADS]: The instructions include the installation of the @membranehq/cli package from the npm registry. This package is the official tool provided by the vendor (membranedev) for this skill's functionality.
[DATA_EXPOSURE]: The skill emphasizes a 'least privilege' approach to credentials by instructing the agent to never ask users for API keys, instead relying on Membrane's server-side authentication management.
[INDIRECT_PROMPT_INJECTION]: As the skill ingests data from external Verimi records and CLI outputs, it possesses an inherent surface for indirect prompt injection.
Ingestion points: Data is received via membrane action run and membrane request commands.
Boundary markers: None explicitly defined in the provided instructions.
Capability inventory: The agent can execute shell commands via the Membrane CLI and perform network requests through the Membrane proxy.
Sanitization: No specific sanitization or validation of the retrieved Verimi data is described.

verimi