quickbooks

Pass

Audited by Gen Agent Trust Hub on Apr 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the membrane CLI tool for all interactions with the QuickBooks API. Commands such as membrane login, membrane connection, and membrane action are invoked to manage the lifecycle of the integration. This is the intended design for skills within the Membrane ecosystem.\n- [EXTERNAL_DOWNLOADS]: During setup, the skill requires the installation of the @membranehq/cli package from the public npm registry. This is a vendor-owned resource (Membrane) and is necessary for the skill's functionality.\n- [PROMPT_INJECTION]: The skill interacts with external data from QuickBooks (e.g., invoices and customer details), which presents a surface for indirect prompt injection. Given the primary purpose of the skill is data integration and the use of structured JSON for outputs, this risk is assessed as safe within the intended use-case.\n
  • Ingestion points: Data retrieved via membrane action run commands.\n
  • Boundary markers: Not explicitly defined in the instructions.\n
  • Capability inventory: Access to the host shell to execute membrane CLI commands.\n
  • Sanitization: Relies on the agent's internal logic for handling command-line arguments and processing JSON responses.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 22, 2026, 05:17 PM