The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @membranehq/cli package via NPM. This is a tool provided by the platform vendor for interacting with their services.
[COMMAND_EXECUTION]: Uses the membrane CLI for managing connections and executing actions. These are legitimate operations within the scope of the skill's purpose.
[DATA_EXFILTRATION]: Authentication is handled by the Membrane CLI's login and connect mechanisms, which avoid direct handling of sensitive API keys or tokens by the agent instructions.
[PROMPT_INJECTION]: The skill processes external data via action inputs, creating a surface for indirect prompt injection.
Ingestion points: membrane action run uses the --input parameter to pass data to actions (SKILL.md).
Boundary markers: None identified in the prompt templates.
Capability inventory: membrane action run and membrane action create (SKILL.md) allow interaction with external APIs.
Sanitization: Not explicitly defined in the skill instructions.

klaro-app