strategy-generate

What problem does it solve?

It turns high-level trading intents into structured strategy workflows, parsing instruments, time ranges, and signal logic so agents can confidently design and evaluate models without missing critical details like normalization or validation needs.

Core Features & Use Cases

• Requirement parsing: normalize instrument codes, infer or request missing dates, and capture entry/exit logic before writing config.json.
• Signal design and coding: follow the five-question checklist, implement SignalEngine in pandas, and ensure boundary-safe, normalized signal series.
• Evaluation loop: leverage the built-in backtest tool, examine artifacts/metrics, and iterate on edits until trade_count, equity, and review criteria gain confidence; ideal for dual MA, RSI, or crypto trend ideas.

Quick Start

Ask the agent to run strategy-generate with your instrument, date range, and objectives so it can write config.json, craft signal_engine.py, and launch the built-in backtest.