Borrowing it
Nothing to install: this file belongs to aliveranme/stata-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/aliveranme/stata-mcp/master/CLAUDE.mdgit clone --depth 1 https://github.com/aliveranme/stata-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/aliveranme/stata-mcp/claude-md)<a href="https://agentmods.dev/instructions/aliveranme/stata-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/aliveranme/stata-mcp/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/aliveranme/stata-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/aliveranme/stata-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.24127 | $0.24127 |
| Opus 5 | $0.12063 | $0.12063 |
| Sonnet 5 | $0.04825 | $0.04825 |
| Haiku 4.5 | $0.02413 | $0.02413 |
Grade A, and why
stata-mcp CLAUDE.md scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
| 慢网络下 `setup.py` 裸 traceback 退出 | 五处 `subprocess.run` 都传了 timeout 却无一捕获 `TimeoutExpired`（它继承自 Exception 而非 OSError） | `install_deps` 捕获并给出重试与手动安装命令；uv 路径超时改走 pip 回退 | How it starts
The opening of the file, as written. The whole thing — 592 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Stata MCP Server
让 Claude Code Agent 通过 MCP Server 直接驱动 Stata，自动完成数据加载、清洗、建模、结果导出全流程。
架构
stata-mcp/
├── mcp-stata-server/server.py # MCP 执行层：75 个工具，通过 pystata 调用 Stata DLL
├── mcp-stata-server/tool_modules/ # 便利工具模块（数据重构/扩展估计/后估计），register() 装配
├── .claude/skills/stata/SKILL.md # 知识层：Stata 语法、模板、陷阱、Agent 协作规范
├── setup.py # 安装层：检测 Stata、创建 venv、生成 .mcp.json
├── .gitignore # 忽略 .mcp.json(生成) .venv dta/log/smcl
关键链路：Claude + SKILL.md → 调用 MCP 工具 → server.py → pystata → Stata DLL
会话持久：Stata 在 MCP Server 启动时初始化一次，所有工具调用共享同一会话，数据跨调用保持。
命令
| 命令 | 说明 |
|---|---|
python setup.py |
一键安装（检测 Stata → venv → fastmcp → .mcp.json → 验证） |
cd mcp-stata-server && source .venv/Scripts/activate && python server.py |
调试模式启动 |
uv pip install fastmcp && uv pip freeze > requirements.txt |
添加新依赖 |
MCP 工具（75 个）
能力边界不在工具数上：
stata_run执行任意命令、stata_help查任意命令的 官方语法，二者即「全量内置命令支持」。专用工具（回归/面板/IV/生成变量等）是 给高频命令加结构化参数与校验的便利层，不是能力上限。
| 类别 | 工具 | 只读? | 说明 |
|---|---|---|---|
| 核心执行 | stata_run, stata_run_do_file |
— | 通用命令执行；run_do_file 执行前自动拆出 ssc install 单独安装（已装跳过）。stata_run 另有 save_output：完整输出（不受 120K 裁剪）落盘并登记为文件资源 |
| 数据管理 | stata_use_dataset, stata_import, stata_save_dataset, stata_set_cwd |
— | 读写 .dta、cd；stata_import 覆盖官方 import 族（excel/delimited/sas/spss/dbase/parquet，按扩展名推断）。save_dataset 成功后自动登记为资源 |
| 面板/时序 | stata_xtset |
— | 声明/查询/清除 xtset(面板) 与 tsset(纯时序) —— 是 stata_xtreg 的前提 |
| 示例数据 | stata_use_example |
— | sysuse(本地) / webuse(联网) 加载官方示例数据集；action="list" 列出可用 |
| 数据生成 | stata_generate, stata_egen |
— | 创建变量（改数据集，非只读）；支持官方 [type] 存储类型与 [if] [in] |
| 数据重构 | stata_merge, stata_append, stata_reshape, stata_collapse, stata_frame, stata_replace, stata_drop, stata_keep, stata_rename, stata_recode, stata_destring |
— | 横向合并(1:1/m:1/1:m/m:m)、纵向追加(可多文件)、长宽转换、按组聚合、多数据集 frame；后六个是变量级清洗：replace/recode 覆盖原变量，drop/keep 支持「删变量」或「删观测」两种形态（二选一），destring 强制 replace 或 generate() 二选一 |
| 数据校验 | stata_verify |
✓ | count/assert/duplicates/isid/missing 五合一（missing 走 misstable summarize） |
| 数据探索 | stata_describe, stata_codebook, stata_summarize, stata_list, stata_tabulate, stata_correlate, stata_display |
✓ | 只读探索 |
| 估计 | stata_regress, stata_logistic, stata_probit, stata_poisson, stata_ttest, stata_xtreg, stata_ivregress, stata_logit, stata_mlogit, stata_nbreg, stata_qreg, stata_mixed |
✓ | OLS/Logit/Probit/Poisson/t 检验/面板/IV + 扩展族：logit（原始系数，logistic 是 OR）、mlogit 多分类、nbreg 负二项、qreg 分位（quantile 默认 0.5）、mixed 多水平（random 以 || 开头） |
| 后估计 | stata_margins, stata_test, stata_predict, stata_estat, stata_estimates, stata_lincom, stata_nlcom, stata_hausman |
✓* | 边际效应/Wald 检验/预测（stata_predict 会创建变量，非只读）；stata_estat 诊断(vif/hettest/ovtest/ic)；stata_estimates 存取与并排比较模型；lincom 线性组合、nlcom 非线性组合（delta 法）、hausman 模型比较（需先 estimates store 两个模型） |
| 返回值 | stata_return_list |
✓ | 一次列出 r()/e()/c() 全部返回值，不必逐个 display |
| 图形 | stata_graph, stata_scheme |
— / ✓* | 绘图并可选导出（选项按格式自动适配，见下）；stata_scheme 列出/查询/设置主题（action="set" 非只读）。stata_graph 导出成功后自动登记为资源 |
| 导出 | stata_export_excel, stata_export_delimited, stata_etable |
— | 数据集导出为 .xlsx 或 CSV/TSV/自定义分隔符（replace 默认 False）；回归表导出优先用 stata_etable（官方 etable，无第三方依赖，直出 .docx/.xlsx/.pdf/.tex）。export_excel(results=True) 是旧路径：依赖第三方 estout 且只能产出 CSV。导出成功即登记为资源 |
| 文件资源回传 | stata_read_file, stata_register_file, stata_list_resources |
✓ / — | 导出工具成功后会登记输出文件，远程客户端可经 MCP 资源协议（resources/read 读 stata-file:///<路径>）或 stata_read_file（base64）取回图表/Excel/CSV/dta 的实际内容，而不只是路径。安全边界：只读登记过的文件，未登记报错并提示登记方式 |
| 包管理与帮助 | stata_install_package, stata_uninstall_package, stata_describe_package, stata_find_package, stata_list_packages, stata_help |
— / ✓ | 装/卸（ado uninstall 本地安全）/查详情（本地 ado describe 或联网 ssc describe）/net search 找包/ado dir 列包/stata_help 查任意命令帮助 |
| 会话生命周期 | stata_clear, stata_snapshot |
— | clear 按 scope 重置（data/estimates/graphs/panels/all）；snapshot 包 Stata 原生快照 save/list/restore/erase，同会话内数据阶段间快速回退 |
| 长任务控制 | stata_background, stata_task_status, stata_task_cancel, stata_task_result, stata_task_list |
— / ✓ | 后台执行长任务（大循环/复杂回归/联网，单块最长 3600s），立即返回任务号；进度轮询、显式取消、取结果。后台任务仍持 _stata_lock，运行期间其他调用会等待 |
| 翻页 | stata_more |
✓ | 大输出分页浏览（缓存 120K chars） |
| 会话 | stata_status |
✓ | 数据集 + 工作目录 + frame + 面板/时序设定 + 已存/活跃估计 + 内存 —— 覆盖 Agent 调 xtreg/margins/predict 前需确认的全部前提 |
| 心跳 | stata_ping |
✓ | 快速检测 Stata DLL 存活状态 |
| 服务器日志 | stata_read_log |
✓ | 读取本 MCP Server 的运行日志（tail/path），排查远程客户端看不到的服务器侧问题 |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 592 lines · 24,127 tokens per session scan A 9de1d1502d62
stata-mcp CLAUDE.md is an instructions file published in the GitHub repository aliveranme/stata-mcp (0 stars, last pushed 10d ago), licensed MIT. It adds 24,127 tokens to every session, about $0.1206 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.