The Agent Skills Directory

[SAFE]: The skill is functionally aligned with its stated purpose of systematic trading. It implements multiple layers of checks, including a backtest gate and human-in-the-loop approval processes.- [COMMAND_EXECUTION]: The skill is designed to place financial orders via Robinhood. This behavior is documented as the primary goal and includes hard caps for risk management that the skill claims are deterministic and outside the LLM context.- [EXTERNAL_DOWNLOADS]: The skill references data ingestion from well-known and trusted financial data providers, specifically Alpaca and Yahoo Finance (yfinance). These references are standard for the financial domain.- [PROMPT_INJECTION]: The skill processes market data from external sources (Alpaca/yfinance). While this represents a potential surface for indirect prompt injection (Category 8), the risk is mitigated by the 'notification-first' human approval step and the requirement for strategies to pass a separate backtest gate before being used in execution.

stock-daytrading