The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @membranehq/cli global Node.js package. This is a vendor-owned resource used to facilitate communication with the Membrane platform.\n- [COMMAND_EXECUTION]: The skill uses the Membrane CLI to perform operations like authentication (membrane login), connecting to services (membrane connect), and running actions. These are legitimate administrative operations for the service.\n- [DATA_EXFILTRATION]: The skill provides access to actions that can retrieve sensitive data, such as get-current-api-key. This is a documented feature of the integration platform; however, users should be aware that the agent has the capability to access these credentials through the provided tools.\n- [PROMPT_INJECTION]: The skill processes untrusted data from external API responses, which creates an indirect prompt injection surface. Ingestion points: Data enters the agent context through the output of the membrane action run command. Boundary markers: None identified; there are no explicit instructions to ignore or delimit instructions within action outputs. Capability inventory: The agent can execute arbitrary shell commands via the Membrane CLI and create new actions using membrane action create. Sanitization: No sanitization or validation of the ingested external content is mentioned.

openrouter