Zeyang Strategy Lab Free
Use this skill for single-asset daily historical research with the shared automatic data layer. Without --input, the standard entrypoint resolves a supported US ticker or China A-share code, fetches daily history, normalizes it, validates it, and writes a data manifest. An explicit --input remains available for local, offline, or custom research.
Supported strategy shape
- EMA or SMA: close above the selected average means long; close below means cash.
- RSI: below a declared lower threshold means long; above a declared upper threshold means cash.
- Long/cash only, with signal at close and execution on the following bar.
Prefer to parse the symbol, daily timeframe, rule, execution assumption, and cost assumptions from the user's request. Ask the user only when a required field is genuinely missing or the strategy has a material ambiguity. If no input file is supplied, use the shared data layer and preserve its provider, adjustment, fallback, and hash metadata. If a local CSV is supplied, it requires date and close; open is optional. With the default next_open execution, a missing open formally falls back to the next bar's close and produces a warning. Return the generated Strategy Contract, data manifest, summary, core metrics, warnings, and artifact paths. State that the result is research only and not investment advice.
Unsupported requests
Do not search for Pro files, call a Pro runtime, implement an unsupported feature, silently simplify a strategy, or fall back to another rule.
For MACD, Bollinger, cross/crossunder, shorting, leverage, minute data, multi-asset portfolios, or parameter optimization, respond exactly:
Current Free edition does not support this strategy.
Data and interpretation limits
Read references/data-contract.md before accepting ambiguous or malformed input. For metric definitions and research limits, read references/metrics-and-robustness.md. Do not claim that execution lag proves an external signal has no look-ahead bias.
