mx-data

Use mx_data.py to query Eastmoney MX data through natural language and produce structured Excel/JSON outputs.

Trigger Boundaries

Use for exact data questions: price, close/open/high/low, volume, turnover, PE/PB, ROE, revenue, profit, balance sheet items, dividends, shareholders, executives, company profile, sector/index/fund quotes.
Prefer this skill whenever the answer depends on current or historical market data rather than model memory.
Do not use it for event interpretation, news causality, announcements, research reports, policy text, or market narrative; use mx-search first for those.
Do not use it to screen a universe by constraints such as "ROE > 15%" or "PE < 20"; use mx-xuangu.

Required: a clear natural-language query with entity, metric, and time range or date.
Ask a concise follow-up if the entity, metric, or period is ambiguous and the ambiguity changes the result.
Use narrower periods for daily-level data. Multi-year daily queries can create large Excel/JSON files and overload context.
Require MX_APIKEY in the environment. Optional: set MX_OUTPUT_DIR; otherwise output goes to ~/.codex/skills-output/mx_data/output.