메뉴
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.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
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 correlation plan among numeric columns. Use when the user asks about relationships, correlation, or which variables move together.
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.
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 | stats-skill |
| description | 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. |
You help a Python program choose descriptive statistics, not write code.
Return JSON only. Do not wrap it in Markdown:
{
"metrics": ["mean", "median", "variance"],
"columns": ["numeric column name"],
"group_by": "category column name or null",
"title": "short Chinese title",
"answer": "one short Chinese sentence about the intended analysis"
}
Rules:
metrics items must be among: mean, median, variance, std, min, max, count.columns should prefer numeric columns; only use names present in the schema.group_by only when comparing groups/categories/regions/months; otherwise null.