The Agent Skills Directory

[SAFE]: The skill integrates with Qwilr, a document automation tool, via the Membrane platform. It follows standard integration practices for this environment.
[EXTERNAL_DOWNLOADS]: The skill guides the user to install the Membrane CLI (@membranehq/cli) via npm. This is a vendor-owned tool required for the skill's operation.
[COMMAND_EXECUTION]: The skill uses the membrane CLI for platform operations such as login, connect, and action run. These commands are used to interact with the vendor's service and are expected for this integration.
[DATA_EXPOSURE_AND_EXFILTRATION]: No evidence of sensitive data exposure or unauthorized exfiltration was found. The skill specifically advises letting Membrane handle credentials server-side to avoid local secret storage.
[INDIRECT_PROMPT_INJECTION]: The skill includes an interface for creating actions based on natural language descriptions (membrane action create "DESCRIPTION"). This constitutes a data ingestion surface where external input influences action generation.
Ingestion points: The DESCRIPTION parameter in membrane action create.
Boundary markers: Not explicitly present in the command template.
Capability inventory: Action creation and execution via the Membrane CLI.
Sanitization: Handled server-side by the Membrane platform.

qwilr