The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the @membranehq/cli package globally via npm, which is the official command-line tool for the vendor's platform.
[COMMAND_EXECUTION]: The skill executes various membrane CLI commands to handle authentication (login), establish connections (connect), and manage actions (action list, action create, action run).
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by processing data returned from external API actions.
Ingestion points: Data entering the agent context via the output of the membrane action run command in SKILL.md.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential instructions embedded in the action output.
Capability inventory: The skill allows the agent to create and run arbitrary actions through the CLI, providing a path for multi-step execution chains.
Sanitization: There is no description of sanitization, validation, or escaping of action output before it is processed by the agent.

loopmessage