e-goi

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the @membranehq/cli package globally via npm, which is the official tool for the Membrane platform provided by the vendor.
  • [COMMAND_EXECUTION]: Shell commands are used to manage authentication, connection state, and action execution through the CLI interface.
  • [PROMPT_INJECTION]: The skill incorporates an interface for discovering and creating actions based on natural language input (e.g., membrane action list --intent). This represents a surface for indirect prompt injection where user-supplied queries influence the platform's tool discovery, though it is a core feature of the service's orchestration model.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 02:18 PM