### Install Stata Packages: Examples

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/plugins/stata-toolbox/skills/stata-skill/references/ado_package_install.md

Demonstrates installing packages from SSC, GitHub, and a custom URL. Includes examples for specifying replacement and source.

```python
# Install outreg2 from SSC
ado_package_install("outreg2")

# Install from GitHub without overwriting
ado_package_install("user/repo", source="github", is_replace=false)

# Install from a custom URL
ado_package_install("custompkg", source="net", package_source_from="https://site.example/stata")
```

--------------------------------

### Install Dependencies and Setup Development Environment

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/CLAUDE.md

Installs project dependencies using uv and sets up the package in development mode. Verifies the installation by checking the version.

```bash
# Install dependencies and create virtual environment
uv sync

# Install the package in development mode
uv pip install -e .

# Verify installation
stata-mcp --version
```

--------------------------------

### Quick Start: Core API Functions

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/api/overview.md

Demonstrates basic usage of key MCP-for-Stata API functions including getting data info, installing packages, executing do-files, and reading log files.

```python
from stata_mcp.api import (
    stata_do,
    get_data_info,
    ado_package_install,
    read_log,
    stata_help,
    write_dofile,
)

# Get data information
info = get_data_info("/path/to/data.dta")

# Install an enabled and approved package
result = ado_package_install("outreg2", source="ssc")

# Execute a do-file
log = stata_do("/path/to/analysis.do")

# Read log file
content = read_log("/path/to/output.log")
```

--------------------------------

### Quick Start: Use Stata Agent as a Tool

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/agent_as/agent_as_tool.md

Initialize a StataAgent and convert it into a tool for use with a main Agent. This example assumes default OpenAI models.

```python
import asyncio

from agents import Agent
from stata_mcp.agent_as.agent_as_tool import StataAgent

# init stata agent and set as tool
stata_agent = StataAgent()
sa_tool = stata_agent.as_tool()

# Create main Agent
agent = Agent(
    ...,
    tools=[sa_tool],
)


# Then run the agent as usual.
async def main():
    ...


if __name__ == "__main__":
    asyncio.run(main())
```

--------------------------------

### Run MCP-for-Stata Commands with uvx

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/cli.md

Examples of using uvx to manage and run MCP-for-Stata commands without a global installation. This includes checking the version, running diagnostics, executing do-files, starting the server, and installing to a client.

```bash
uvx stata-mcp --version
```

```bash
uvx stata-mcp doctor
```

```bash
uvx stata-mcp tool do /path/to/analysis.do
```

```bash
uvx stata-mcp server
```

```bash
uvx stata-mcp install -c cursor
```

--------------------------------

### Install MCP-for-Stata Manually

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/agents/openclaw.md

Recommended installation using uvx or manual installation via pip.

```bash
# Run directly with uvx (recommended)
uvx stata-mcp tool --help

# Or install and use
upip install stata-mcp
stata-mcp tool --help
```

--------------------------------

### Install All Supported Clients with MCP

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/clients.md

Automatically configure all supported AI clients by running `stata-mcp install --all`.

```bash
stata-mcp install --all
```

--------------------------------

### Install and Configure Node.js with nvm

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs Node.js version 20.19.5, sets it as the default, and makes it available for use. Also checks the installed versions of Node.js and npm.

```bash
nvm install 20.19.5  # At least you should install 18+
nvm alias default 20
nvm use default
node -v
npm -v
```

--------------------------------

### Quick Setup Environment Variables

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/configuration.md

Common environment variables for quick setup, including debug mode, working directory, monitoring, and logging.

```bash
# Enable debug mode
export STATA_MCP__IS_DEBUG=true

# Set working directory
export STATA_MCP__CWD="/projects/my-analysis"

# Enable monitoring with 8GB RAM limit
export STATA_MCP__IS_MONITOR=true
export STATA_MCP__RAM_LIMIT=8192

# Disable security guard (not recommended)
export STATA_MCP__IS_GUARD=false

# Enable console logging
export STATA_MCP__LOGGING_CONSOLE_HANDLER_ON=true
```

--------------------------------

### Install MCP-for-Stata using Beta Script (macOS/Linux)

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/README.md

Installs MCP-for-Stata and automatically installs `uv` if it's missing. Run this command in your terminal.

```bash
curl -fsSL https://raw.githubusercontent.com/SepineTam/mcp-for-stata/master/scripts/install.sh | bash
```

--------------------------------

### Install to Specific Client

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/cli.md

Installs MCP-for-Stata to a single, specified AI client using the '-c' or '--client' flag. Examples include 'claude-code' and 'cursor'.

```bash
# Install to a specific client
stata-mcp install -c claude-code
stata-mcp install -c cursor
```

--------------------------------

### Install uvx from PyPI with pipx

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs uvx from the Python Package Index (PyPI) using pipx, which is recommended for isolated installations.

```bash
pipx install uv
```

--------------------------------

### Install uvx from astral.sh

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs uvx using a script provided by astral.sh. This method downloads and executes an installation script directly.

```bash
curl -LsSf https://astral.sh/uv/install.sh | sh
```

--------------------------------

### Install Stata-MCP and Verify Installation

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/agent_examples/README.md

Install the Stata-MCP package using pip and verify the installation by checking the version.

```bash
git clone https://github.com/sepinetam/stata-mcp.git
cd stata-mcp

pip install -e .
# You can find whether it is installed successfully by:
stata-mcp --version
```

--------------------------------

### Install MCP-for-Stata using Beta Script (Windows PowerShell)

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/README.md

Installs MCP-for-Stata and automatically installs `uv` if it's missing. Run this command in PowerShell.

```powershell
irm https://raw.githubusercontent.com/SepineTam/mcp-for-stata/master/scripts/install.ps1 | iex
```

--------------------------------

### Install uvx with Homebrew

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs uvx, a Python version manager, using Homebrew. This is an alternative method to installing uvx.

```bash
brew install uvx
uvx --version
```

--------------------------------

### Minimal PromptGenerator Example

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/agent_examples/task_prompt/README.md

A minimal example demonstrating initialization with default parameters and generating both instructions and tasks. Data paths can be specified using natural language.

```python
from prompt_generator import PromptGenerator

generator = PromptGenerator()

instructions = generator.instructions()
tasks = generator.tasks(
    datas="use Stata default data",  # you can use natural language to set teh paths of data.
    aims="Explore the relationship between a and b"
)

print(instructions)
print(tasks)
```

--------------------------------

### Execute Self-contained Install Scripts

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/cli.md

Demonstrates how to run the self-contained installation scripts for various platforms (Unix shells, macOS Finder, Windows PowerShell, Windows command line). These scripts bootstrap a Python toolchain and install stata-mcp.

```bash
bash install.sh
```

```bash
powershell -ExecutionPolicy Bypass -File install.ps1
```

--------------------------------

### One-Click Installation for All Agents

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/README.md

Installs MCP-for-Stata for all supported agents with a single command. No manual configuration is required.

```bash
uvx stata-mcp install --all
```

--------------------------------

### Install uv Package Runner

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/claude-plugin.md

Install the 'uv' package runner, recommended for managing Python packages. Installation can be done via Homebrew or the official installer script.

```bash
# Install via homebrew
brew install uv
```

```bash
# Or via official installer
curl -LsSf https://astral.sh/uv/install.sh | sh
```

--------------------------------

### Install MCP-for-Stata to Specific Client

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/plugins/stata-toolbox/skills/stata-skill/references/installation.md

Install MCP-for-Stata to a specific AI client by specifying the client flag.

```bash
uvx stata-mcp install -c codex
```

```bash
uvx stata-mcp install -c claude
```

```bash
uvx stata-mcp install -c cc
```

```bash
uvx stata-mcp install -c cursor
```

```bash
uvx stata-mcp install -c cline
```

```bash
uvx stata-mcp install -c gemini
```

```bash
uvx stata-mcp install -c opencode
```

```bash
uvx stata-mcp install -c openclaw
```

```bash
uvx stata-mcp install -c hermes
```

--------------------------------

### Install Matplotlib

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/reports/ai_coding_ability/main.ipynb

Install the matplotlib library if you intend to create figures. This command uses 'uv pip' for installation.

```python
# if you want to make a figure about it, you should install matplotlib.
# !uv pip install matplotlib
```

--------------------------------

### Install stata-mcp

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/agent_as/agent_as_tool.md

Install the stata-mcp package using pip or uv.

```bash
pip install stata-mcp
```

```bash
uv add stata-mcp
```

--------------------------------

### RAM Limit Environment Variable Examples

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/monitoring.md

Examples of setting the RAM limit using environment variables, including no limit, a specific limit, and recommended values.

```bash
# No limit (default)
export STATA_MCP__RAM_LIMIT=-1

# 4 GB limit
export STATA_MCP__RAM_LIMIT=4096

# 8 GB limit
export STATA_MCP__RAM_LIMIT=8192

# 16 GB limit
export STATA_MCP__RAM_LIMIT=16384
```

--------------------------------

### Install Single Client with MCP

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/clients.md

Use `stata-mcp install` to automatically configure a specified client. Replace `<client>` with a valid client key.

```bash
stata-mcp install -c <client>
```

--------------------------------

### Add and Install Stata MCP Plugin

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/claude-plugin.md

Add the Stata MCP plugin marketplace and install the 'stata-toolbox' plugin. The plugin can be installed to user, project, or local scope.

```bash
# Add marketplace
claude plugin marketplace add sepinetam/stata-mcp
```

```bash
# Install plugin to user scope (default)
claude plugin install stata-toolbox
```

```bash
# Install to project scope (shared with team)
claude plugin install stata-toolbox --scope project
```

```bash
# Install to local scope (gitignored)
claude plugin install stata-toolbox --scope local
```

--------------------------------

### Install Stata Package from Network Source

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/agents/openclaw.md

Installs a Stata package from a network URL. Use --source net and provide the --package-source-from URL.

```bash
# Install from network
stata-mcp tool ado-install custompkg --source net --package-source-from "https://example.com/stata/" --yes
```

--------------------------------

### Install uvx from PyPI with pip

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs uvx directly using pip. Note that a Rust toolchain might be required if a prebuilt wheel is not available for your platform.

```bash
pip install uv
```

--------------------------------

### Install Stata Package from GitHub

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/agents/openclaw.md

Installs a Stata package from a GitHub repository. Specify --source github and the repository name.

```bash
# Install from GitHub
stata-mcp tool ado-install SepineTam/TexIV --source github --yes
```

--------------------------------

### Install Stata MCP Plugin

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/usage.md

Installs the stata-toolbox plugin for MCP server and Stata LSP configuration.

```bash
claude plugin install stata-toolbox@stata-plugin-lib -s project
```

--------------------------------

### List Installed MCP Servers

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/usage.md

Verifies the installation of MCP servers by listing them.

```bash
claude mcp list
```

--------------------------------

### Install uvx from astral.sh with wget

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs uvx using a script from astral.sh, utilizing wget if curl is not available. This is an alternative to the curl method.

```bash
wget -qO- https://astral.sh/uv/install.sh | sh
```

--------------------------------

### Install stata-skill via OpenClaw

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/index.md

Use this command to install the stata-skill from ClawHub. This is part of the OpenClaw integration.

```text
Install `stata-skill` from ClawHub by @SepineTam. 
```

--------------------------------

### Install Stata Language Server (LSP)

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/claude-plugin.md

Install the 'stata-language-server' globally using npm if it's not detected or if you need to ensure the latest version is available.

```bash
npm install -g stata-language-server
```

--------------------------------

### Install psutil

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/monitoring.md

Ensure the psutil library is installed and meets the minimum version requirement (>= 6.0.0) for Stata MCP monitoring.

```bash
uv pip install psutil>=6.0.0
```

--------------------------------

### Start MCP Server with Tool Profile Selection

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/CLAUDE.md

Starts the Stata-MCP server, allowing selection of tool profiles (all tools or core tools) and transport methods.

```bash
# Start with tool profile selection
stata-mcp server                     # All tools, stdio (same as bare command)
stata-mcp server --core              # Core tools only (stata_do, get_data_info, help)
stata-mcp server --all -t http       # All tools, HTTP transport
stata-mcp server --core -t http      # Core tools, HTTP transport
```

--------------------------------

### Install Stata Package

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/agents/openclaw.md

Install an enabled and approved Stata package using the `ado-install` tool. The `--yes` flag bypasses confirmation prompts.

```bash
stata-mcp tool ado-install outreg2 --yes
```

--------------------------------

### Install Claude Plugin for MCP-for-Stata

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/usage.md

Add the MCP-for-Stata marketplace registry and install the plugin globally for use with Claude Code.

```bash
# Add marketplace registry
claude plugin marketplace add sepinetam/stata-mcp

# Install plugin globally
claude plugin install stata-toolbox -s user
```

--------------------------------

### Install Stata Package from SSC

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/agents/openclaw.md

Installs an approved Stata package from the SSC archive. Use the --yes flag to skip interactive confirmation.

```bash
# Install an approved SSC package
stata-mcp tool ado-install outreg2 --yes
```

--------------------------------

### Install to All Clients Explicitly

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/cli.md

Installs MCP-for-Stata to all supported clients by explicitly using the '--all' flag. This is functionally equivalent to omitting the '-c' and '--json-file' flags.

```bash
# Install to all clients explicitly
stata-mcp install --all
```

--------------------------------

### Operational Examples for Help Resource

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/mcp/resources.md

Demonstrates how to use the help resource to fetch documentation for various Stata commands across different categories.

```python
# Regression commands
help("regress")
help("logit")
help("probit")

# Panel data commands
help("xtset")
help("xtreg")
help("xtmixed")

# Data management
help("merge")
help("reshape")
help("collapse")

# Time series
help("tsset")
help("arima")
help("varsoc")
```

--------------------------------

### Install to Custom JSON Config

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/cli.md

Installs MCP-for-Stata into a custom JSON configuration file specified by '--json-file'. This allows for managing installations in non-default locations or custom setups.

```bash
# Install into a custom JSON config file
stata-mcp install --json-file /path/to/config.json
```

--------------------------------

### Basic Monitoring Setup

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/monitoring.md

Enable monitoring using environment variables and run MCP-for-Stata. Monitoring is automatic and requires no code changes.

```bash
export STATA_MCP__IS_MONITOR=true
export STATA_MCP__RAM_LIMIT=8192

stata-mcp
# or
stata-mcp agent run
```

--------------------------------

### Allowlist GitHub Repositories for Ado Installation

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/configuration.md

Configure the security settings to allow specific GitHub repositories for ado package installations. This requires adding approved repositories to the allowlist and starting the MCP server with the --unsafe flag.

```toml
[SECURITY]
ADO_INSTALL_ALLOWED_GITHUB_REPOSITORIES = ["SepineTam/TexIV"]
```

--------------------------------

### Add MCP-for-Stata Globally for Claude Code

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/README.md

Installs MCP-for-Stata globally for the current user, making it available for use with Claude Code. This is a one-time setup.

```bash
claude mcp add stata-mcp --scope user -- uvx stata-mcp
```

--------------------------------

### Install MCP-for-Stata Skill

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/skills/openclaw.md

Download the skill's Markdown file to a specified directory. This is a prerequisite for using the skill.

```bash
mkdir -p /path/to/skill/base/stata-mcp-skill
curl -o /path/to/skill/base/stata-mcp-skill/SKILL.md https://github.com/SepineTam/mcp-for-stata/raw/refs/heads/master/docs/agents/openclaw.md
```

--------------------------------

### Install Stata-MCP in Claude Code

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/00_intro.md

Use this command to add the stata-mcp package within the Claude Code environment. Ensure you have the necessary permissions and environment setup.

```bash
claude mcp add stata-mcp uvx stata-mcp
```

--------------------------------

### Install Client with Custom JSON Configuration

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/clients.md

Configure a client using a custom JSON file, optionally specifying a nested JSON key path with `--json-index`.

```bash
stata-mcp install -c <client> --json-file /path/to/config.json
stata-mcp install -c <client> --json-file /path/to/config.json --json-index parent.child
```

--------------------------------

### Plugin Configuration Example

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/claude-plugin.md

Specifies the metadata and server configurations for a plugin package. Includes version, author, repository, and server definitions.

```json
{
  "name": "stata-toolbox",
  "version": "0.1.0",
  "description": "The official working package of MCP-for-Stata plugin, including mcp config and stata lsp.",
  "author": {
    "name": "Song Tan",
    "email": "sepinetam@gmail.com",
    "url": "https://www.sepinetam.com"
  },
  "homepage": "https://statamcp.com",
  "repository": "https://github.com/sepinetam/mcp-for-stata",
  "license": "AGPL-3.0",
  "keywords": ["stata", "econometrics", "empirical analysis"],
  "mcpServers": {
    ...
  },
  "lspServers": {
    ...
  }
}
```

--------------------------------

### Example TOML Configuration File

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/configuration.md

A sample configuration file demonstrating various settings for debugging, logging, help caching, security, project directories, and Stata-specific options.

```toml
[DEBUG]
IS_DEBUG = false

[DEBUG.logging]
LOGGING_ON = true
LOGGING_CONSOLE_HANDLER_ON = false
LOGGING_FILE_HANDLER_ON = true
LOG_FILE = "~/.statamcp/stata_mcp_debug.log"
MAX_BYTES = 10_000_000
BACKUP_COUNT = 5

[BETA]
ENABLE_WRITE_DOFILE = false

[HELP]
IS_CACHE = true
IS_SAVE = true

[SECURITY]
IS_GUARD = true
ADO_INSTALL_ALLOWED_GITHUB_REPOSITORIES = []

[PROJECT]
WORKING_DIR = ""
CLEAN_LOG_DAYS = -1
FOLDER_TAG = ".statamcp"

[MONITOR]
IS_MONITOR = false
MAX_RAM_MB = -1

[STATA]
# Optional: Override automatic Stata detection
# STATA_CLI = "/path/to/stata-mp"

[data_info]
metrics = ["obs", "mean", "stderr", "min", "max", "q1", "q3", "skewness", "kurtosis"]
string_keep_number = 10
decimal_places = 3
hash_length = 12

```

--------------------------------

### Install Claude Code with npm

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs Claude Code globally using npm. This is the primary method for installing the tool.

```bash
npm install -g @anthropic-ai/claude-code
```

--------------------------------

### Install Homebrew

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs Homebrew, a package manager for macOS, using a bash script. This is a prerequisite for installing other development tools.

```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
```

--------------------------------

### Configure Unified Settings via Environment Variables

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/usage.md

Quickly set up MCP-for-Stata configurations, including working directory, security guard, RAM monitoring, and RAM limit, using environment variables.

```bash
# Quick setup with environment variables
export STATA_MCP__CWD="/projects/my-analysis"
export STATA_MCP__IS_GUARD=true
export STATA_MCP__IS_MONITOR=true
export STATA_MCP__RAM_LIMIT=16384
```

--------------------------------

### Install MCP-for-Stata with Codex CLI

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/README.md

Use this command to install MCP-for-Stata via the Codex CLI. Ensure you have uv installed.

```bash
uvx stata-mcp install -c codex
```

--------------------------------

### Install Claude Code

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/claude-plugin.md

Install the Claude Code application using the official installer script. This is a prerequisite for using the Stata plugin.

```bash
# macOS/Linux
curl -fsSL https://claude.ai/install.sh | bash
```

```powershell
# Windows
irm https://claude.ai/install.ps1 | iex
```

--------------------------------

### TOML Environment Configuration with All Features

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/clients.md

Example of a TOML environment configuration enabling all security, monitoring, and logging features.

```toml
env.STATA_MCP__CWD = "/path/to/project"
env.STATA_MCP__IS_GUARD = "true"
env.STATA_MCP__IS_MONITOR = "true"
env.STATA_MCP__RAM_LIMIT = "8192"
env.STATA_MCP__LOGGING_CONSOLE_HANDLER_ON = "true"
```

--------------------------------

### Manual Agent Setup for Evaluation

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/Evaluation.md

An alternative method for setting up an agent manually using the OpenAI client if you prefer not to use AgentRunner directly. Requires manual extraction of data.

```python
# If you prefer to set up the agent manually
from openai import OpenAI
from agents import Agent, Runner

client = OpenAI(api_key="your-api-key")
agent = Agent(
    instructions="You are a helpful assistant specialized in Stata analysis.",
    model="gpt-3.5-turbo"
)

result = client.agent.run(agent, input=YOUR_TASK)
# Then extract data manually as needed
```

--------------------------------

### Install Stata Package without Replacing Existing

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/agents/openclaw.md

Installs a Stata package but prevents replacement of existing files. Useful for checking installation status.

```bash
# Don't replace existing package (useful for checking installation status)
stata-mcp tool ado-install estout --yes --is-replace false
```

--------------------------------

### List Installed Claude Plugins

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/claude-plugin.md

Verify that the Stata plugin is loaded by listing all installed Claude plugins. This command confirms the successful installation and availability of the plugin.

```bash
# List installed plugins
claude plugin list
```

--------------------------------

### Marketplace Manifest Example

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/claude-plugin.md

Defines the structure for a plugin library registry. Includes library name, owner details, and a list of plugins.

```json
{
  "name": "stata-plugin-lib",
  "owner": {
    "name": "Song Tan",
    "email": "sepinetam@gmail.com"
  },
  "plugins": [
    {
      "name": "stata-toolbox",
      "source": "./plugins/stata-toolbox",
      "description": "The official working package of MCP-for-Stata plugin, including mcp config and stata lsp."
    }
  ]
}
```

--------------------------------

### Install Stata Package from SSC

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/plugins/stata-toolbox/skills/stata-skill/references/ado_package_install.md

Installs a Stata package from the SSC archive. This is the default source. Consider checking if the package is already installed to avoid unnecessary downloads.

```python
ado_package_install("outreg2")
ado_package_install("reghdfe")
ado_package_install("estout", is_replace=false)
```

--------------------------------

### Programmatic RAM Monitoring Setup

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/monitoring.md

Instantiate RAMMonitor with a specific limit and pass it to StataDo for programmatic monitoring.

```python
from stata_mcp.monitor import RAMMonitor
from stata_mcp.core.stata import StataDo

# Create monitor with 8GB limit
monitor = RAMMonitor(max_ram_mb=8192)

# Pass to StataDo
stata = StataDo(
    dofile_path="analysis.do",
    monitors=[monitor]  # Optional: list of monitors
)

# Execution is automatically monitored
result = stata.execute()
```

--------------------------------

### Verify Installation

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/cli.md

Check if the stata-mcp CLI is installed and accessible by displaying its version.

```bash
stata-mcp --version
```

--------------------------------

### Install Stata Package

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/mcp/tools.md

Installs a Stata package from various sources like SSC, GitHub, or a network URL. Use this to manage Stata package installations programmatically. Ensure the 'unsafe' profile is enabled for this tool.

```python
def ado_package_install(package: str,
                        source: str = "ssc",
                        is_replace: bool = False,
                        package_source_from: str | None = None) -> str:
    ...
```

```python
# SSC package installation
ado_package_install("outreg2", source="ssc")
```

```python
# GitHub package installation
ado_package_install("sepinetam/texiv", source="github")
```

```python
# Network installation
ado_package_install("custompkg", source="net", package_source_from="https://example.com/stata")
```

```python
# Force reinstall
ado_package_install("estout", source="ssc", is_replace=True)
```

--------------------------------

### Install Stata Package from GitHub

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/plugins/stata-toolbox/skills/stata-skill/references/ado_package_install.md

Installs a Stata package from a GitHub repository. The repository must follow Stata package conventions and be listed in the allowlist. GitHub packages do not receive security protection, so inspect them before installation.

```python
ado_package_install("SepineTam/TexIV", source="github")
```

--------------------------------

### Add New Configuration Option

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/CONTRIBUTING.md

Example of how to implement a new configuration option in the project's configuration system. This involves defining a property that retrieves a configuration value, potentially from environment variables, with a default and a validator.

```python
MY_NEW_OPTION:
    return self._get_config_value(
        config_keys=["SECTION", "MY_OPTION"],
        env_var="STATA_MCP__MY_OPTION",
        default=False,
        converter=self._to_bool,
        validator=lambda x: isinstance(x, bool)
    )
```

--------------------------------

### Project-based MCP Server Installation

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/usage.md

Installs the stata-mcp MCP server scoped to a specific project directory.

```bash
cd ~/Documents/MyResearchProject
claude mcp add stata-mcp --env STATA_MCP__CWD=$(pwd) --scope project -- uvx --directory $(pwd) stata-mcp
```

--------------------------------

### JSON Environment Configuration with Security and Monitoring

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/clients.md

Example of a JSON environment configuration enabling security guard and RAM monitoring.

```json
"env": {
  "STATA_MCP__CWD": "/path/to/project",
  "STATA_MCP__IS_GUARD": "true",
  "STATA_MCP__IS_MONITOR": "true",
  "STATA_MCP__RAM_LIMIT": "8192"
}
```

--------------------------------

### Create and Execute a Stata Do-file

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/api/overview.md

Demonstrates how to write a Stata do-file programmatically and then execute it using the stata_do function. Ensure the do-file path is correctly specified.

```python
dofile_path = write_dofile("""
use \"/data/survey.dta\", clear
regress income age education
estat hettest
""")

# Execute it
result = stata_do(dofile_path)
```

--------------------------------

### Start MCP Server with Stdout/Stdin Transport

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/CLAUDE.md

Starts the Stata-MCP server using the default stdio transport.

```bash
# Start MCP server with stdio transport (default)
stata-mcp
```

--------------------------------

### Install to All Supported Clients

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/cli.md

Installs MCP-for-Stata to all supported AI clients without specifying individual clients or a JSON file. This is the default behavior when no client or JSON file options are provided.

```bash
# Install to all supported clients (no -c, no --json-file)
stata-mcp install
```

--------------------------------

### Install Claude Code with Homebrew (Beta)

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/source/docs/Usages/ClaudeCode/01_install.md

Installs Claude Code using Homebrew. This method is noted as being in beta.

```bash
brew install --cask claude-code
```

--------------------------------

### get_data_info Operational Examples

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/mcp/tools.md

Demonstrates various ways to call the get_data_info function, including local files, remote URLs, and specific encodings.

```python
# Local file analysis
get_data_info("/data/econometrics/survey.dta")
get_data_info("~/Documents/exports/quarterly.csv", vars_list=["gdp", "inflation", "unemployment"])
```

```python
# Remote data ingestion
get_data_info("https://repository.org/datasets/panel_data.xlsx")
```

```python
# Encoded source handling
get_data_info("/data/legacy/latin1_data.csv", encoding="latin1")
```

--------------------------------

### Install SSC Package

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/plugins/stata-toolbox/skills/stata-skill/SKILL.md

Requests approval to install a package from the SSC archive. Ensure the package name is correct.

```python
ado_package_install("outreg2")
```

--------------------------------

### Troubleshooting Configuration Loading: Check Config File

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/configuration.md

Verify the existence of the TOML configuration file in the default location.

```bash
ls ~/.statamcp/config.toml
```

--------------------------------

### Install MCP-for-Stata Skills in OpenClaw

Source: https://github.com/sepinetam/mcp-for-stata/blob/master/docs/agents/openclaw.md

Use this message to instruct OpenClaw to install MCP-for-Stata skills from the provided URL.

```text
Install MCP-for-Stata Skills from https://docs.statamcp.com/skills/openclaw
```