The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the membrane CLI tool for operations such as connection management, action discovery, and running API requests.
[EXTERNAL_DOWNLOADS]: The skill requires the global installation of the @membranehq/cli npm package, which is a verified resource provided by the author.
[REMOTE_CODE_EXECUTION]: The skill suggests using npx @membranehq/cli@latest, which dynamically downloads and executes the vendor's CLI tool from the npm registry.
[PROMPT_INJECTION]: The skill processes potentially untrusted content from Reward Gateway (e.g., employee comments, announcements, and articles), which introduces a surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context through the output of membrane action run and membrane request commands.
Boundary markers: There are no explicit boundary markers or instructions to treat external data as untrusted in the skill's logic.
Capability inventory: The skill has the capability to execute CLI commands and interact with external systems through the Membrane proxy.
Sanitization: The skill does not provide mechanisms for sanitizing or validating API responses before they are processed by the agent.

reward-gateway