deepgram

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill prompts the user to install the '@membranehq/cli' package from NPM, which is the official tool for the Membrane platform.
  • [COMMAND_EXECUTION]: The instructions involve executing various 'membrane' shell commands to perform authentication, create connections, and run API actions.
  • [SAFE]: The skill demonstrates a high security posture by leveraging Membrane's managed connection system, which prevents the need for manual API key handling or exposure of secrets in the user environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 09:05 PM