data
GitHub该技能用于数据诊断,通过获取schema和样本评估数据结构、质量及可分析性。涵盖表规模、字段类型、缺失值、异常检测等,区分事实与建议,明确可直接分析的字段及需确认口径。
Trigger Scenarios
需要查看数据结构或概览时
评估数据质量和可分析性时
Install
npx skills add Zafer-Liu/Data-Analysis-Agent --skill data -g -y
SKILL.md
Frontmatter
{
"icon": "🔍",
"name": "data",
"description": "查看数据结构、质量和可分析性(schema 数据概览 profiling)",
"allowedTools": [
"get_schema",
"get_table_detail",
"query_data",
"profile_data",
"generate_chart"
]
}
数据诊断
先获取 schema 和必要样本,报告表规模、字段类型、缺失、重复、异常、时间范围和关键分布。区分事实、风险与建议,指出可直接分析的字段及需要用户确认的口径。
Tool routing
- Use
get_schemafirst to inspect tables, fields, row counts, and source structure. - Use
get_table_detailwhen one table needs deeper field-level metadata. - Use
profile_datafor data quality, missingness, type, and distribution diagnostics. - Use
query_datafor small verification samples or targeted aggregates. - Use
generate_chartonly when a compact diagnostic chart materially helps explain the data.
Implementation reference
- Data tool entries:
agent/tools/business/data.py - Profiling implementation:
Function/Clean/data_profile.py - Chart implementation:
Function/Charts_generation/chart_generate.py
Version History
- d6a2c3e Current 2026-07-24 12:12


