The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the membrane CLI to interact with Podio data, perform searches, and manage actions, which is the intended functional behavior for this integration.\n- [EXTERNAL_DOWNLOADS]: The instructions guide the installation of the @membranehq/cli package from the official NPM registry; this is a trusted vendor resource.\n- [REMOTE_CODE_EXECUTION]: The membrane action create command is used to dynamically build integration logic on the Membrane platform based on natural language descriptions, which is a core platform capability.\n- [CREDENTIALS_UNSAFE]: The skill demonstrates safe credential handling by instructing the agent to use the platform's connection manager rather than requesting or storing raw API keys or tokens.\n- [PROMPT_INJECTION]: Indirect injection surface analyzed: 1. Ingestion points: User-provided intent and description fields in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: Execution of membrane CLI commands via shell. 4. Sanitization: Not specified in instructions. This is a functional surface required for natural language tool discovery and is assessed as safe.

podio