agentleFS
LambenthanLambenthan/empiricalwiki/.claude/skills/stata-plan/SKILL.md

stata-plan

Skill85 starsChanged 5 months agoMITView on GitHub

将实证研究设计转成 Stata 执行计划或 do 文件骨架,包含数据合并、变量构造、描述统计、回归和稳健性

Files
SKILL.md

/stata-plan

把 wiki 中的变量、数据和模型信息转成 Stata 层面的执行顺序。默认生成计划;只有用户传 --write-do 时才创建 .do 骨架。

Workflow

Step 1: Read Design and Data Context

读取:

  • 指定的 wiki/outputs/empirical-design-*.md,或按 topic 匹配最近设计文档
  • wiki/variables/*.md
  • wiki/datasets/*.md
  • wiki/models/*.md
  • wiki/robustness/*.md
  • wiki/heterogeneity/*.md
  • wiki/identification/*.md(设计涉及 DID / PSM / IV / RDD / TWFE 时优先读对应预置策略卡)
  • tools/stata-templates/*.do(预置模板:twfe / did_staggered / psm / iv_2sls / rdd;--write-do 时以匹配的模板为骨架起点,按设计文档改写变量与参数)
  • 项目中的 .do、.dta、.xlsx、变量字典和 README

Step 2: Produce Stata Execution Plan

输出结构:

# Stata 执行计划

## 1. 输入数据
## 2. 主键与合并顺序
## 3. 样本筛选
## 4. 变量构造
## 5. 缩尾与缺失值处理
## 6. 描述性统计
## 7. 相关性分析
## 8. 基准回归
## 9. 机制检验
## 10. 异质性检验
## 11. 稳健性检验
## 12. 表格导出
## 13. 复核清单

每一步尽量写成 Stata 伪代码或代码块,但不要假造不存在的变量名。变量名不确定时用占位符并标记“需确认”。

Step 3: Optional Do File

只有用户传 --write-do 时,创建:

wiki/outputs/stata-plan-{slug}-{YYYY-MM-DD}.do

默认只写:

wiki/outputs/stata-plan-{slug}-{YYYY-MM-DD}.md

Step 4: Log

python3 tools/research_wiki.py log wiki "stata-plan | <topic> | output: outputs/<file>"

Constraints

  • 不覆盖用户已有 .do 文件。
  • 不假设 Stata 包已经安装;需要 reghdfe、esttab 等时在计划里列出安装命令。
  • 代码骨架必须保留复核点,如样本量检查、merge 结果检查、变量分布检查。
  • 对中文路径和空格路径使用引号。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.