The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the installation of the @membranehq/cli package from the NPM registry. This is an official tool provided by the vendor (membranedev/membrane) and is a standard dependency for accessing the platform.
[COMMAND_EXECUTION]: The skill makes extensive use of the membrane CLI to perform operations such as authentication, action discovery, and running specific integration tasks. These commands are part of the intended system interaction.
[CREDENTIALS_UNSAFE]: The instructions demonstrate good security hygiene by explicitly telling the agent to avoid requesting or handling sensitive user credentials directly, delegating this to the Membrane platform's server-side connection manager.
[PROMPT_INJECTION]: The skill processes data returned from external Insightsoftware actions, which creates a surface for indirect prompt injection. Attackers with control over data in the connected system could potentially embed instructions aimed at manipulating the agent's behavior.
Ingestion points: Data returned from action executions (e.g., reports, user lists) is processed by the agent.
Boundary markers: None identified; the skill does not specify the use of delimiters or 'ignore' instructions for the external data.
Capability inventory: The CLI provides capabilities for file system interactions, network requests, and the generation of new executable actions.
Sanitization: There are no instructions for sanitizing or validating the contents of the retrieved data before the agent acts upon it.

insightsoftware