菜单
Convert a data question into a correlation plan among numeric columns. Use when the user asks about relationships, correlation, or which variables move together.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Convert a data question into a dataset-profile plan (overview, schema, or quality). Use when the user asks what the dataset looks like, what columns/types exist, how many rows, or about data quality and missing values.
Convert a data question into a distribution plan for one numeric column (histogram or box plot). Use when the user asks about distribution, spread, outliers, or the shape of a single numeric variable.
Convert a data question into a time-series trend plan. Use when the user asks how a numeric value changes over time, monthly/weekly/daily trends, or seasonality.
Convert a data question into a visualization plan (bar/line/scatter/pie) for tabular data. Use when the user asks to compare, rank, trend by category, relate two numbers, or visualize a breakdown.
Convert a data question into a descriptive-statistics plan for tabular data. Use when the user asks for mean, median, variance, std, min, max, count, or grouped descriptive statistics.
快速验证 Flask API 端点可用性，自动检测改动影响的接口并跑测试
| name | correlation-skill |
| description | Convert a data question into a correlation plan among numeric columns. Use when the user asks about relationships, correlation, or which variables move together. |
You help a Python program plan a correlation analysis, not write code.
Return JSON only. Do not wrap it in Markdown:
{
"columns": ["numeric column names, 2 or more"],
"method": "pearson | spearman | kendall",
"title": "short Chinese title",
"answer": "one short Chinese sentence"
}
Rules: