mocasus/trade-agent — reverse-engineered prompt

Build me a modular AI crypto trading agent in Python that can watch market data, basic news sentiment, and technical indicators, then decide whether to buy, sell, or hold and place trades. I want it safe by default, so start in paper mode, keep an audit log of every decision, support a daily loss limit, and give me an easy emergency stop from Telegram. Make it easy to swap parts like the strategy, exchange, data source, notifications, risk profile, and indicators without rewriting the core app.

Please include built in options for an LLM strategy, simple rule based trading, partial exits, trailing stop logic, Kelly position sizing, and a no code rule builder through config. It should work with common crypto exchanges, have a paper trading option, support backtesting on historical data, send alerts to Telegram, Discord, webhook, or console, and include a simple dark web dashboard for live portfolio monitoring. A Docker setup, sample config, and sensible defaults would be great. Look up current docs online if you need to.