human-api

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the Membrane CLI (membrane login, membrane action run, membrane connect) to manage health data. It also includes membrane action create, which allows for dynamic creation of API interactions based on natural language descriptions.
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of a global Node.js package (@membranehq/cli@latest) from the NPM registry. This package is the official tool for the platform described in the skill and is owned by the vendor.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection as it processes health data from external sources and creates actions dynamically.
  • Ingestion points: Data retrieved from Human API through membrane action run and metadata from membrane action list.
  • Boundary markers: None provided in the command templates.
  • Capability inventory: Subprocess execution via the CLI, network access to the Human API, and dynamic logic generation via the action creation feature.
  • Sanitization: No specific sanitization or filtering instructions are provided for the data retrieved from external sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 07:37 PM
Security Audit — agent-trust-hub — human-api