The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the @membranehq/cli package globally using npm. This is a vendor-provided tool necessary for the skill's functionality.
[COMMAND_EXECUTION]: It uses the membrane CLI to perform authentication, manage platform connections, and execute remote actions. This includes dynamically creating actions based on natural language descriptions.
[PROMPT_INJECTION]: The skill processes external data from BotConversa (e.g., messages and subscriber info) which represents a surface for Indirect Prompt Injection.
Ingestion points: Output from membrane action run in SKILL.md.
Boundary markers: Absent; there are no specific delimiters used to wrap external content or instructions for the agent to ignore instructions within that content.
Capability inventory: Subprocess execution via membrane CLI in SKILL.md, which includes capabilities to write data back to the external platform.
Sanitization: Absent; the skill does not specify any validation or filtering of the data retrieved from the external API.

botconversa