herald

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @membranehq/cli package from the official NPM registry. This is a vendor-provided tool (Membrane) necessary for the skill's core functionality.
  • [COMMAND_EXECUTION]: The skill relies on the membrane CLI to perform operations such as logging in, connecting to services, and running actions. These are legitimate uses of the vendor's command-line interface for platform integration.
  • [DATA_EXFILTRATION]: The skill demonstrates positive security practices by explicitly instructing the agent to never ask the user for API keys or tokens, relying instead on Membrane's server-side connection management to handle the authentication lifecycle securely.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 08:43 AM