alpaca-trading-mcp-server

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly lists and enables a "news" toolset and gives the example "Get the latest news about Tesla", and the MCP server can act on market data/news (via trading tools) so the agent will fetch untrusted third-party news/content through Alpaca and may let that content influence trading actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The Docker deployment instructions explicitly git clone https://github.com/alpacahq/alpaca-mcp-server.git and then build/run the image, meaning remote code from that repository is fetched and executed as a required MCP server at runtime.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill is explicitly a trading integration (Alpaca MCP Server) built to execute real trades and manage brokerage accounts. The documentation requires Alpaca API keys, supports switching to live trading, and documents placing market/limit/stop orders, submitting/canceling/modifying orders, closing positions, exercising options, and managing account balances. Code examples show submit_order/MarketOrderRequest and live trading modes. These are specific financial execution APIs (market orders, crypto trades, options), not generic tooling—so it grants direct financial execution authority.