The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of @membranehq/cli from the global NPM registry. This is a verified vendor tool provided by the author to facilitate secure interactions with their integration platform.
[COMMAND_EXECUTION]: All documented shell commands are restricted to the membrane CLI tool. These commands are used for standard lifecycle operations such as authentication, service connection, and action execution.
[DATA_EXFILTRATION]: Security is enhanced by utilizing the platform's centralized credential management. This approach avoids exposing sensitive API tokens within the skill's scripts or local environment.
[PROMPT_INJECTION]: The skill includes an indirect prompt injection surface as it processes external data via CLI arguments.
Ingestion points: User-supplied input flows into the membrane action list and membrane action run commands in SKILL.md via the --intent and --input flags.
Boundary markers: None identified in the provided documentation.
Capability inventory: The skill can execute CLI commands and interact with the Impression API through the membrane binary.
Sanitization: No explicit sanitization or validation of the input parameters is described in the skill instructions.

impression