robinhood-connector

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: Integration with the official Robinhood Agentic MCP endpoint. The skill references https://agent.robinhood.com/mcp/trading, which is a service provided by a well-known and established financial institution.
  • [SAFE]: Secure default operating mode. The skill is designed to operate in 'notification mode' by default, meaning the agent only proposes orders for human review and approval rather than executing trades autonomously.
  • [SAFE]: Explicit safety gating. Transitioning to live trading requires multiple manual steps, including setting environment variables and daily human confirmation, which prevents accidental or unauthorized automated trading.
  • [SAFE]: Code-side safety invariants. The skill documentation describes the implementation of hard caps and a kill-switch mechanism (.KILL file) that operate outside the influence of the language model.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 08:02 AM
Security Audit — agent-trust-hub — robinhood-connector