The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the membrane CLI to manage connections, list actions, and execute workflows. These commands are part of the standard operating procedure for the Membrane platform and do not involve unauthorized privilege escalation.- [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @membranehq/cli package globally via npm. This is the official command-line tool provided by the vendor (Membrane) to facilitate the integration and is considered a trusted source.- [PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection where user-supplied intents, descriptions, and input parameters are passed to the Membrane platform.
Ingestion points: intent in membrane action list, DESCRIPTION in membrane action create, and input parameters in membrane action run.
Capability inventory: The tool allows running actions and creating new ones on the BugBug service.
Boundary markers/Sanitization: Not explicitly defined within the skill instructions, though the skill mandates the use of Membrane's managed connection system to handle credentials safely, reducing the risk of data exfiltration.

bugbug